AI Assets API provides endpoints to retrieve, update, and create several types of AI assets, such as systems, data sets, prompts, and models.

The AI Assets API supports actions that you can perform on AI Asset records in the Expanded Model and Asset Classes application. It requires the Asset Classes (sn_ent) plugin to access it. You must have the asset and model_manager roles to call the endpoints provided by the AI Assets API.
Note: The data included in AI Asset records may be spread across several tables like Product Model, Configuration Item, and others. To reduce complexity when calling this API, however, the AI Assets API inserts data into only the Asset, Product Model and Configuration Item tables. Specifics about table updates are provided under each endpoint.

AI Assets API - GET /sn_ent/asset/ai_dataset/{sys_id}

Retrieves the data of the specified AI dataset asset.

URL format

Versioned URL: /api/sn_ent/{api_version}/asset/ai_dataset/{sys_id}

Default URL: /api/sn_ent/asset/ai_dataset/{sys_id}

Supported request parameters

Table 1. Path parameters
Name Description
api_version Optional. Version of the endpoint to access. For example, v1 or v2. Only specify this value to use an endpoint version other than the latest.

Data type: String
sys_id Sys_id of the asset in the AI Dataset Asset [alm_ai_dataset_digital_asset] table.

Data type: String
Table 2. Query parameters
Name Description
None
Table 3. Request body parameters (XML or JSON)
Name Description
None

Headers

The following request and response headers apply to this HTTP action only, or apply to this action in a distinct way. For a list of general headers used in the REST API, see Supported REST API headers.

Table 4. Request headers
Header Description
Accept Data format of the response body. Only supports application/json.
Content-Type Data format of the request body. Only supports application/json.
Table 5. Response headers
Header Description
None

Status codes

The following status codes apply to this HTTP action. For a list of possible status codes used in the REST API, see REST API HTTP response codes.

Table 6. Status codes
Status code Description
200 Successful. The request was successfully processed.
401 Unauthorized. The user credentials are incorrect or have not been passed.
404 Not found. Failed to fetch the asset with the given sys_id.
500 Internal server error. An unexpected error occurred while processing the request. The response contains additional information about the error.

Response body parameters

Name Description
result Details of the newly created AI Dataset asset.

Data type: Object

"result": {
  "acceptable_usage": {Object},
  "base_datasets": [Array],
  "created": "String",
  "dataset_card": "String",
  "data_type": {Object},
  "description": "String",
  "documentation": "String",
  "display_name": "String",
  "managed_by": {Object},
  "name": "String",
  "provider": {Object},
  "state": "Development",
  "source": "String",
  "sys_id": "String",
  "updated": "String",
  "version": "String",
}
result.acceptable_usage Acceptable usage for the AI Dataset Asset record. Acceptable usage refers to how a dataset or model can be used, typically for training or evaluation purposes.

Data type: Object

"acceptable_usage": {
  "label": "String" 
  "value": "String" 
}
result.acceptable_usage.label Display label of the acceptable usage value.

Data type: String
result.acceptable_usage.value Number value of the acceptable usage.
Valid values:
  • 1: Training
  • 2: Evaluation

Data type: String
result.base_datasets Comma-separated list of base datasets required to build the given dataset. Accepts the name or sys_id of a base data set in the AI Dataset Digital Asset [alm_ai_dataset_digital_asset] table.

Data type: Array

"base_datasets": ["String", "String"]
result.created Date and time that the AI Dataset Asset record was created.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.data_type The type of data present in the AI Dataset Asset record.

Data type: Object

"data_type": {
  "label": "String",
  "value": "String"
}
result.data_type.label The display label for the data type value.

Data type: String
result.data_type.value Value of the dataset asset's data type.

Data type: String
result.dataset_card The data set card. A dataset_card is a metadata document that describes the contents, structure, and context of an AI dataset. It provides details like data sources, features, intended use, and any known limitations to ensure proper understanding and usage.

Data type: String
result.description Description of the associated AI Dataset Product Model record.

Table: AI Dataset Product Model [cmdb_ai_dataset_product_model]

Data type: String
result.display_name Display name of the AI Dataset Asset record.

Table: AI Dataset Digital Asset [alm_ai_dataset_digital_asset]

Data type: String
result.documentation Documentation of the associated AI Dataset Product Model [cmdb_ai_dataset_product_model] table record.

Data type: String
result.managed_by Details about the user who manages the AI Dataset Asset record.

Data type: Object

"managed_by": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.managed_by.name Name of the user who manages the AI Dataset Asset record.

Table: User [user]

Data type: String
result.managed_by.sys_id Sys_id of the user who manages the AI Dataset Asset record.

Table: User [user]

Data type: String
result.name Name of the associated AI Dataset Product Model record.

Table: AI Dataset Product Model [cmdb_ai_dataset_product_model]

Data type: String
result.provider Provider of the associated AI Dataset Product Model [cmdb_ai_dataset_product_model] table record.

Data type: Object

provider: {
  "name": "String",
  "sys_id": "String"
}
result.provider.name Name of the provider.

Data type: String
result.provider.sys_id Sys_id of the provider in the associated AI Dataset Product Model [cmdb_ai_dataset_product_model] table record.

Data type: String
result.source Details about the source of AI dataset asset.
Valid values:
  • Link to the source of the dataset asset.
  • Details (in plain text) of the source of the dataset asset. For example, the name of a product or website.

Data type: String

Default: empty or null
result.state State of the AI Dataset Asset record.
Possible values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String
result.sys_id Sys_id of the AI Dataset Asset record.

Table: AI Dataset Digital Asset [alm_ai_dataset_digital_asset]

Data type: String
result.updated Date and time that the AI Dataset Asset record was last updated.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.version Version number of the associated AI Dataset Product Model record. For example, V2.

Data type: String

Example: cURL request

The following example.

curl "https://instance.servicenow.com/api/sn_ent/asset/ai_dataset/ cc419cb2331e92101c9aca989d5c7b4c" \ 
--request GET \ 
--header "Accept:application/json" \ 
--user "username":"password"

Output:

{ 
  "result": { 
    "sys_id": "9d60fb5f40d21210f877b00c113d1fea", 
    "display_name": "ServiceNow Closed Incidents v1", 
    "name": "Closed Incidents", 
    "description": "Incidents with resolution", 
    "documentation": "Sample Documentation", 
    "source": "incident table on servicenow instance", 
    "dataset_card": "Sample Dataset Card", 
    "state": "Development", 
    "version": "v1", 
    "data_type": { 
      "value": "2", 
      "label": "Image" 
    }, 
    "provider": { 
      "sys_id": "93d4ecfac0a8000b6294d71b733977fb", 
      "name": "ServiceNow" 
    }, 
    "managed_by": { 
      "sys_id": "62826bf03710200044e0bfc8bcbe5df1", 
      "name": "Abel Tuter" 
    }, 
    "acceptable_usage": { 
      "value": "1", 
      "label": "Training" 
    }, 
    "base_datasets": [], 
    "created": "2024-12-11 08:50:40", 
    "updated": "2024-12-11 08:50:40" 
  } 
}

AI Assets API - GET /sn_ent/asset/ai_model/{sys_id}

Retrieves the data of the specified AI model asset.

URL format

Versioned URL: /api/sn_ent/{api_version}/asset/ai_model/{sys_id}

Default URL: /api/sn_ent/asset/ai_model/{sys_id}

Supported request parameters

Table 7. Path parameters
Name Description
api_version Optional. Version of the endpoint to access. For example, v1 or v2. Only specify this value to use an endpoint version other than the latest.

Data type: String
sys_id Sys_id of the asset in the AI Model Asset [alm_ai_model_digital_asset] table.

Data type: String
Table 8. Query parameters
Name Description
None
Table 9. Request body parameters (XML or JSON)
Name Description
None

Headers

The following request and response headers apply to this HTTP action only, or apply to this action in a distinct way. For a list of general headers used in the REST API, see Supported REST API headers.

Table 10. Request headers
Header Description
Accept Data format of the response body. Only supports application/json.
Content-Type Data format of the request body. Only supports application/json.
Table 11. Response headers
Header Description
None

Status codes

The following status codes apply to this HTTP action. For a list of possible status codes used in the REST API, see REST API HTTP response codes.

Table 12. Status codes
Status code Description
200 Successful. The request was successfully processed.
401 Unauthorized. The user credentials are incorrect or have not been passed.
404 Not found. Failed to fetch the asset with the given sys_id.
500 Internal server error. An unexpected error occurred while processing the request. The response contains additional information about the error.

Response body parameters
result Details of the model asset record.

Data type: Object

result: {
  "base_model": {Object},
  "context_window": "String",
  "created": "String",
  "deployment_guideline": "String",
  "description": "String",
  "display_name": "String",
  "documentation": "String",
  "evaluation_datasets": [Array],
  "evaluation_metrics_report": "String",
  "managed_by": {Object},
  "model_size_in_mb": "String",
  "name": "String",
  "parameters_info": "String",
  "provider": {Object},
  "required_infrastructure": "String",
  "source": "String",
  "state": Number,
  "supported_languages": [Array],
  "sys_id": "String",
  "training_datasets": [Array],
  "training_procedure": "String",
  "updated": "String",
  "version": "String"
}
result.base_model AI model that this model version was derived from.
Note: Only applicable for models developed within the organization.

Data type: Object

{
  "name": "String",
  "sys_id": "String"
 }
Default:
result.base_model.name Name of the AI model asset to model this AI model after.

Data type: String
result.base_model.sys_id Sys_id of the AI model asset to model this AI model after.

Data type: String
result.context_window Size of input sequences (in other words, the number of tokens) that the model can handle.

Data type: String represented with a number
result.created Date and time that the AI Dataset Asset record was created.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.deployment_guideline Instructions applicable for models developed and deployed within an organization.

Data type: String
result.description Description to give the AI Model Product Model.

Updated in table: AI Model Product Model [cmdb_ai_model_product_model]

Data type: String
result.result.display_name Display name of the asset record.

Data type: String
result.documentation Documentation of the AI Prompt Product Model record.

Table: AI Prompt Product Model [cmdb_ai_model_product_model]

Data type: String
result.evaluation_datasets Comma-separated list of sys_ids or display names of AI datasets of the AI Dataset Digital Asset used for evaluating the model. Mostly applicable for models developed within an organization.

Tables: AI Dataset Digital Asset [alm_ai_dataset_digital_asset]

Data type: Array

"evaluation_datasets": [
  "name": "String",
  "sys_id": "String"
]
result.evaluation_metrics_report Reference to the evaluation results.
Possible values:
  • Details (in plain text) outlining results
  • Links to specific results

Data type: String
result.managed_by Details about the user who manages the AI Model Asset record.

Data type: Object

"managed_by": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.managed_by.name Name of the user who manages the AI Dataset Asset record.

Table: User [user]

Data type: String
result.managed_by.sys_id Sys_id of the user who manages the AI Dataset Asset record.

Table: User [user]

Data type: String
result.model_size_in_mb Size of the model in MB. Mostly applicable for models developed and deployed within an organization.

Data type: Number
result.name Required. Name of the AI Model Product Model.

Updated in table: AI Model Product Model [cmdb_ai_model_product_model]

Data type: String
result.provider Provider of the associated AI Dataset Product Model [cmdb_ai_dataset_product_model] table record.

Data type: Object

provider: {
  "name": "String",
  "sys_id": "String"
}
result.provider.sys_id Sys_id of the provider in the associated AI Dataset Product Model [cmdb_ai_dataset_product_model] table record.

Data type: String
result.provider.name Name of the provider.

Data type: String
result.required_infrastructure Documentation of infrastructure needs for the model deployment. For example, details about the infrastructure stack and processing needs. Mostly applicable for models deployed within an organization.

Data type: String
result.source Details about the source of the model.
Possible values:
  • Link to the source of the model. For example, https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1 or a link to Github.
  • Details of the source in plain text. For example, Microsoft Azure

Data type: String
result.state State to apply to the AI Model Asset.
Possible values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String
result.supported_languages Languages that are supported by the AI Model.

Data type: Array

"supported_languages": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.supported_languages.name Name of the Language record. For example, "French", "English".

Table: Language [sys_language]

Data type: String
result.supported_languages.sys_id Sys_id of the Language record.

Table: Language [sys_language]

Data type: String
result.sys_id Sys_id of the AI model asset record.

Table: AI Model Asset [alm_ai_model_digital_asset]

Data type: String
result.version Version for AI Model Product Model. For example, V2.

Data type: String

Example: cURL request

The following example.

curl "https://instance.servicenow.com/api/sn_ent/asset/ai_model/cc419cb2331e92101c9aca989d5c7b4c" \
--request GET \
--header "Accept:application/json" \
--user "username":"password"

Output:

{
  "result": {
    "sys_id": "a438d170ff96da10c1fbffffffffffd5",
    "display_name": "ServiceNow Now LLM V6",
    "name": "Now LLM",
    "description": "enables text-to-text like question answering and summarization",
    "version": "V6",
    "provider": {
      "sys_id": "93d4ecfac0a8000b6294d71b733977fb",
      "name": "ServiceNow"
    },
    "documentation": "Now LLM V5 Documentation",
    "parameters_info": "7B",
    "supported_languages": [
      {
        "sys_id": "914493a30f320010e96b0e4fef767e90",
        "name": "English"
      }
    ],
    "model_size_in_mb": "87",
    "deployment_guideline": "Deployed on ServiceNow infrastructure",
    "source": null,
    "training_procedure": "2",
    "context_window": "8000",
    "state": "Deployed",
    "required_infrastructure": "undefined",
    "base_model": {
      "sys_id": null,
      "name": ""
    },
    "evaluation_datasets": [
      {
        "sys_id": "45cb45baff06d610c1fbffffffffffa9",
        "name": "ServiceNow Open Incidents"
      }
    ],
    "training_datasets": [
      {
        "sys_id": "45cb45baff06d610c1fbffffffffffa9",
        "name": "ServiceNow Open Incidents"
      }
    ],
    "evaluation_metrics_report": "Testing results: See files attached to this record",
    "managed_by": {
      "sys_id": "62826bf03710200044e0bfc8bcbe5df1",
      "name": "Abel Tuter"
    },
    "created": "2024-12-03 16:50:53",
    "updated": "2024-12-03 16:50:53"
  }
}

AI Assets API - GET /sn_ent/asset/ai_prompt/{sys_id}

Retrieves the data of the specified AI prompt asset.

See also:

URL format

Versioned URL: /api/sn_ent/{api_version}/asset/ai_prompt/{sys_id}

Default URL: /api/sn_ent/asset/ai_prompt/{sys_id}

Supported request parameters

Table 13. Path parameters
Name Description
api_version Optional. Version of the endpoint to access. For example, v1 or v2. Only specify this value to use an endpoint version other than the latest.

Data type: String
sys_id Sys_id of the asset in the AI Prompt Asset [alm_ai_prompt_digital_asset] table.

Data type: String
Table 14. Query parameters
Name Description
None
Table 15. Request body parameters (XML or JSON)
Name Description
None

Headers

The following request and response headers apply to this HTTP action only, or apply to this action in a distinct way. For a list of general headers used in the REST API, see Supported REST API headers.

Table 16. Request headers
Header Description
Accept Data format of the response body. Only supports application/json.
Content-Type Data format of the request body. Only supports application/json.
Table 17. Response headers
Header Description
None

Status codes

The following status codes apply to this HTTP action. For a list of possible status codes used in the REST API, see REST API HTTP response codes.

Table 18. Status codes
Status code Description
200 Successful. The request was successfully processed.
401 Unauthorized. The user credentials are incorrect or have not been passed.
404 Not found. Failed to fetch the asset with the given sys_id.
500 Internal server error. An unexpected error occurred while processing the request. The response contains additional information about the error.

Response body parameters

Name Description
result Details about the AI Prompt Asset.
result: {
  "ai_model": {Object},
  "created": "String",
  "description": "String",
  "display_name": "String",
  "documentation": "String",
  "managed_by": {Object},
  "name": "String",
  "prompt_info": "String",
  "provider": {Object},
  "state": Number,
  "sys_id": "String",
  "updated": "String",
  "version": "String"
}
result.ai_model AI Model of the AI Prompt Asset record.

Data type: Object

"ai_model": {
  "name": "String",
  "sys_id": "String"
}
result.ai_model.name Name of the AI model.

Data type: String
result.ai_model.sys_id Sys_id of the AI Prompt Product Model record.

Table: AI Prompt Product Model [alm_ai_model_digital_asset]

Data type: String
result.created Date and time that the AI Prompt Asset record was created.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.description Description of the AI Prompt Product Model record.

Table: AI Prompt Product Model [cmdb_ai_prompt_product_model]

Data type: String
result.display_name Display name of the associated AI Prompt Asset record.

Table: AI Prompt Asset [alm_ai_prompt_digital_asset]

Data type: String
result.documentation Documentation for the AI Prompt Product Model record.

Table: AI Prompt Product Model [cmdb_ai_prompt_product_model]

Data type: String
result.managed_by User who manages the AI Prompt Asset record.

Data type: Object

"managed_by": {
  "name": "String",
  "sys_id": "String"
}
result.managed_by.name Name of the user who manages the asset record.

Data type: String
result.managed_by.sys_id Sys_id of the User record that corresponds to the Managed by user of the asset record.

Tables: User [sys_user], AI Prompt Asset [alm_ai_model_digital_asset]

Data type: String
result.name Name of the associated AI Prompt Product Model.

Table: AI Prompt Product Model [cmdb_ai_prompt_product_model]

Data type: String
result.prompt_info Prompt information of the AI Prompt Asset record.

Table: AI Prompt Asset [alm_ai_prompt_digital_asset]

Data type: String
result.provider Provider of the AI Prompt Product Model record.
"provider": {
  "name": "String",
  "sys_id": "String"
}

Data type: Object
result.provider.name Name of the associated AI Prompt Product Model record.

Data type: String
result.provider.sys_id Sys_id of the Company [core_company] table record that corresponds to the Provider of the associated AI Prompt Product Model record.

Tables: Company [core_company], AI Prompt Product Model [cmdb_ai_prompt_product_model]

Data type: String
result.state State of the AI Prompt Asset record.
Possible values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String
result.sys_id Sys_id of the AI Prompt Asset record.

Table: AI Prompt Asset [alm_ai_prompt_digital_asset]

Data type: String
result.updated Date and time that the AI Prompt Asset record was last updated.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.version Version of the associated AI Prompt Product Model record. For example, V2.

Data type: String

Example: cURL request

The following example shows how to retrieve information about a AI Prompt Asset Model using a specific ID.

curl "https://instance.servicenow.com/api/sn_ent/asset/ai_prompt/cc419cb2331e92101c9aca989d5c7b4c" \
--request GET \
--header "Accept:application/json" \
--user "username":"password"

The response body returns details about the specific prompt asset model that was retrieved.

{
  "result": {
    "sys_id": "cc419cb2331e92101c9aca989d5c7b4c",
    "display_name": "ServiceNow ServiceNow Incident Summarization Prompt 4 V8",
    "name": "ServiceNow Incident Summarization Prompt 4",
    "description": "Prompt for Incident Summarization",
    "version": "V8",
    "provider": {
      "sys_id": "93d4ecfac0a8000b6294d71b733977fb",
      "name": "ServiceNow"
    },
    "documentation": "Documentation",
    "state": "Development",
    "ai_model": {
      "sys_id": "a57d0be6eb1e5210aa82fab8bad0cd18",
      "name": "mistral-large"
    },
    "prompt_info": "Provide incident summary using short_decription, state, worknotes",
    "managed_by": {
      "sys_id": "62826bf03710200044e0bfc8bcbe5df1",
      "name": "Abel Tuter"
    },
    "created": "2024-12-09 03:18:46",
    "updated": "2024-12-09 04:26:08"
  }
}

AI Assets API - GET /sn_ent/asset/ai_system/{sys_id}

Retrieves the data of the specified AI system asset.

URL format

Versioned URL: /api/sn_ent/{api_version}/asset/ai_system/{sys_id}

Default URL: /api/sn_ent/asset/ai_system/{sys_id}

Supported request parameters

Table 19. Path parameters
Name Description
api_version Optional. Version of the endpoint to access. For example, v1 or v2. Only specify this value to use an endpoint version other than the latest.

Data type: String
sys_id Sys_id of the asset in the AI System Digital Asset [alm_ai_system_digital_asset] table.

Data type: String
Table 20. Query parameters
Name Description
None
Table 21. Request body parameters (XML or JSON)
Name Description
None

Headers

The following request and response headers apply to this HTTP action only, or apply to this action in a distinct way. For a list of general headers used in the REST API, see Supported REST API headers.

Table 22. Request headers
Header Description
Accept Data format of the response body. Only supports application/json.
Content-Type Data format of the request body. Only supports application/json.
Table 23. Response headers
Header Description
None

Status codes

The following status codes apply to this HTTP action. For a list of possible status codes used in the REST API, see REST API HTTP response codes.

Table 24. Status codes
Status code Description
200 Successful. The request was successfully processed.
401 Unauthorized. The user credentials are incorrect or have not been passed.
404 Not found. Failed to fetch the asset with the given sys_id.
500 Internal server error. An unexpected error occurred while processing the request. The response contains additional information about the error.

Response body parameters

Name Description
result Details about the retrieved AI System Asset.
result: {
  "ai_models": [Array],
  "ai_prompts": [Array],
  "created": "String",
  "description": "String",
  "display_name": "String",
  "documentation": "String",
  "evaluation_datasets": [Array],
  "evaluation_metrics_report": "String",
  "managed_by": {Object},
  "name": "String",
  "provider": {Object},
  "state": Number,
  "sys_id": "String",
  "updated": "String",
  "version": "String"
}
result.ai_models List of AI models in the AI System Digital Asset [alm_ai_system_digital_asset] table record.

Data type: Array

"ai_models": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.ai_models.name Name of the AI System Digital Asset record.

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: String
result.ai_models.sys_id Sys_id of the AI System Digital Asset record.

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: String
result.ai_prompts List of AI Prompts in the AI System Asset record.

Data type: Array

"ai_prompts": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.ai_prompts.name Name of the AI prompt.

Data type: String
result.ai_prompts.sys_id Sys_id of the AI Prompt Digital Asset record.

Table: AI Prompt Digital Asset [alm_ai_prompt_digital_asset]

Data type: String
result.created Date and time that the AI System Asset was created.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.description Description of the associated AI System Product Model record.

Table: AI System Product Model [cmdb_ai_system_product_model]

Data type: String
result.display_name Display name of the AI System Asset record.

Table: AI System Asset [cmdb_ai_ system_asset_model]

Data type: String
result.documentation Documentation for the AI System Product System record.

Table: AI System Product System [cmdb_ai_system_product_system]

Data type: String
result.evaluation_datasets List of AI datasets used for evaluating the model in the AI System Digital Asset record. Mostly applicable for models developed within an organization.

Tables: AI Dataset Digital Asset [alm_ai_dataset_digital_asset], AI System Asset [alm_ai_system_digital_asset]

Data type: Array

"evaluation_datasets": [
  {
  "name": "String",
  "sys_id": "String"
  }
]

Default: empty string
result.evaluation_datasets.name Name of the AI Dataset Digital Asset.

Data type: String
result.evaluation_datasets.sys_id Sys_id of the AI Dataset Digital Asset record.

Table: AI Dataset Digital Asset [alm_ai_dataset_digital_asset]

Data type: String
result.evaluation_metrics_report Evaluation results of the AI system asset record.
Possible values:
  • Details (in plain text) outlining results
  • Links to specific results

Data type: String
result.managed_by Details about the user who manages the AI System Asset record.

Data type: Object

"managed_by": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.managed_by.name Sys_id of the user who manages the AI System Asset record.

Table: User [user]

Data type: String
result.managed_by.sys_id Name of the user who manages the AI System Asset record.

Table: User [user]

Data type: String
result.name Name of the associated AI System Product Model record.

Table: AI System Product Model [cmdb_ai_system_product_model]

Data type: String
result.provider Value of the Provider field in the associated AI System Product Model record.

Table: AI System Product Model [cmdb_ai_ system_product_model]

Data type: Object

provider: {
  "name": "String",
  "sys_id": "String"
}
result.provider.name Name of the provider in the associated AI System Product Model record.

Table: AI System Product Model [cmdb_ai_system_product_model]

Data type: String
result.provider.sys_id Sys_id of the provider in the associated AI System Product Model record.

Table: AI System Product Model [cmdb_ai_system_product_model]

Data type: String
result.state State of the AI System Asset record.
Possible values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: String
result.sys_id Sys_id of the AI System Asset record.

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: String
result.updated Date and time that the AI System Asset was last updated.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.version Version number of the associated AI System Product Model record. For example, V2.

Data type: String

Example: cURL request

The following example shows how to retrieve an AI System Model Asset with a given ID.

curl "https://instance.servicenow.com/api/sn_ent/asset/ai_system/3b140397435a9210a63d00002fb8f2d7" \
--request GET \
--header "Accept:application/json" \
--user "username":"password"

The response body returns details about the given AI System Model Asset.

{
  "result": {
      "sys_id": "3b140397435a9210a63d00002fb8f2d7",
      "display_name": "ServiceNow Incident Summarization V2",
      "name": "Incident Summarization",
      "description": "Incident Summarization Skill",
      "version": "V2",
      "provider": {
        "sys_id": "93d4ecfac0a8000b6294d71b733977fb",
        "name": "ServiceNow"
      },
      "documentation": "Sample Documentation",
      "state": "Deployed",
      "ai_models": [{ 
        "sys_id": "9tgdc7e6eb1e5210aa82fab8bad0cda2", 
        "name": "llm_generic_small" 
      },
      { 
        "sys_id": "7efdc7e6eb1e5210aa82fab8bad0cda2", 
        "name": "mixtral-instruct" 
      }],
      "ai_prompts": [{ 
        "sys_id": "7d7dc7e6eb1e5210aa82fab8bad0cda2", 
        "name": "LLM Prompt" 
      }],
      "evaluation_datasets": [{ 
        "sys_id": "9d7dc7e6eb1e5210aa82fab8bad0cda2", 
        "name": "Base dataset" 
      }],
      "evaluation_metrics_report": "Sample Report",
      "managed_by": {
        "sys_id": "62826bf03710200044e0bfc8bcbe5df1",
        "name": "Abel Tuter"
      },
      "created": "2024-12-11 18:23:09",
      "updated": "2024-12-11 18:23:09"
  }
}

AI Assets API - POST /sn_ent/asset/ai_dataset

Creates a new AI dataset asset entry in the AI Dataset Digital Asset [alm_ai_dataset_digital_asset] and AI Dataset Product Model [cmdb_ai_dataset_product_model] according to details you provide in the request body.

URL format

Versioned URL: /api/sn_ent/{api_version}/asset/ai_dataset

Default URL: /api/sn_ent/asset/ai_dataset

Supported request parameters

Table 25. Path parameters
Name Description
api_version Optional. Version of the endpoint to access. For example, v1 or v2. Only specify this value to use an endpoint version other than the latest.

Data type: String
Table 26. Query parameters
Name Description
None
Table 27. Request body parameters (XML or JSON)
Name Description
{object} 
{
  "acceptable_usage": "String",
  "base_datasets": [Array],
  "dataset_card": "String"
  "data_type": "String",
  "description": "String",
  "documentation": "String",
  "managed_by": "String", 
  "name": "String", 
  "provider": "String",
  "state": "String" 
  "source": "String",
  "version": "String"
}
{object} 
{
  "acceptable_usage": "String",
  "base_datasets": [Array],
  "dataset_card": "String"
  "data_type": "String",
  "description": "String",
  "documentation": "String",
  "managed_by": "String", 
  "name": "String", 
  "provider": "String",
  "state": "String" 
  "source": "String",
  "version": "String"
}
{object}.acceptable_usage Determines how a dataset or model can be used, typically for training or evaluation purposes.
Valid values:
  • 1: Training
  • 2: Evaluation

Data type: String
{object}.base_datasets Comma-separated list of base datasets needed to build this dataset. Accepts names or sys_ids of datasets present in AI Dataset Digital Asset [alm_ai_dataset_digital_asset] table.

Data type: Array

“base_datasets”: [ “String”, “String”]
{object}.data_type Type of data present in the dataset. For example, Text,Video,Image or 1,2.

Data type: String
{object}.dataset_card The data set card. A dataset_card is a metadata document that describes the contents, structure, and context of an AI dataset. It provides details like data sources, features, intended use, and any known limitations to ensure proper understanding and usage.

Data type: String
{object}.description Description of the associated record in the AI Dataset Product Model [cmdb_ai_dataset_product_model] table.

Data type: String
{object}.documentation Documentation for the AI Dataset Product Model.

Data type: String
{object}.managed_by Value of the Managed by field of an existing User [sys_user] table record.
Valid values:
  • Name of the User [sys_user] record
  • Sys_id of the User [sys_user] record

Data type: String
{object}.name Required. Name of the associated record in the AI Dataset Product Model [cmdb_ai_dataset_product_model] table.

Data type: String
{object}.provider Required. Value of the Provider field of an existing record in the Company [core_company] table.
Valid values:
  • Name of the Company [core_company] record
  • Sys_id of the Company [core_company] record

Data type: String
{object}.source Details about the source of the dataset.
Valid values:
  • Link to the source of the dataset.
  • Details of the source in plain text.

Data type: String

Default: empty or null
{object}.state State of the AI dataset asset.
Valid values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String
{object}.version Version number of the associated AI Dataset Product Model record. For example, V2.

Data type: String

Headers

The following request and response headers apply to this HTTP action only, or apply to this action in a distinct way. For a list of general headers used in the REST API, see Supported REST API headers.

Table 28. Request headers
Header Description
Accept Data format of the response body. Only supports application/json.
Content-Type Data format of the request body. Only supports application/json.
Table 29. Response headers
Header Description
None

Status codes

The following status codes apply to this HTTP action. For a list of possible status codes used in the REST API, see REST API HTTP response codes.

Table 30. Status codes
Status code Description
200 Successful. The request was successfully processed.
400 Bad Request. A bad request type or malformed request was detected.
500 Internal server error. An unexpected error occurred while processing the request. The response contains additional information about the error.

Response body parameters (JSON or XML)

Name Description
result Details of the newly created AI Dataset asset.

Data type: Object

"result": {
  "acceptable_usage": {Object},
  "base_datasets": [Array],
  "created": "String",
  "dataset_card": "String",
  "data_type": {Object},
  "description": "String",
  "documentation": "String",
  "display_name": "String",
  "managed_by": {Object},
  "name": "String",
  "provider": {Object},
  "state": "Development",
  "source": "String",
  "sys_id": "String",
  "updated": "String",
  "version": "String",
}
result.acceptable_usage Acceptable usage for the AI Dataset Asset record. Acceptable usage refers to how a dataset or model can be used, typically for training or evaluation purposes.

Data type: Object

"acceptable_usage": {
  "label": "String" 
  "value": "String" 
}
result.acceptable_usage.label Display label of the acceptable usage value.

Data type: String
result.acceptable_usage.value Number value of the acceptable usage.
Valid values:
  • 1: Training
  • 2: Evaluation

Data type: String
result.base_datasets Comma-separated list of base datasets required to build the given dataset. Accepts the name or sys_id of a base data set in the AI Dataset Digital Asset [alm_ai_dataset_digital_asset] table.

Data type: Array

"base_datasets": ["String", "String"]
result.created Date and time that the AI Dataset Asset record was created.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.data_type The type of data present in the AI Dataset Asset record.

Data type: Object

"data_type": {
  "label": "String",
  "value": "String"
}
result.data_type.label The display label for the data type value.

Data type: String
result.data_type.value Value of the dataset asset's data type.

Data type: String
result.dataset_card The data set card. A dataset_card is a metadata document that describes the contents, structure, and context of an AI dataset. It provides details like data sources, features, intended use, and any known limitations to ensure proper understanding and usage.

Data type: String
result.description Description of the associated AI Dataset Product Model record.

Table: AI Dataset Product Model [cmdb_ai_dataset_product_model]

Data type: String
result.display_name Display name of the AI Dataset Asset record.

Table: AI Dataset Digital Asset [alm_ai_dataset_digital_asset]

Data type: String
result.documentation Documentation of the associated AI Dataset Product Model [cmdb_ai_dataset_product_model] table record.

Data type: String
result.managed_by Details about the user who manages the AI Dataset Asset record.

Data type: Object

"managed_by": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.managed_by.name Name of the user who manages the AI Dataset Asset record.

Table: User [user]

Data type: String
result.managed_by.sys_id Sys_id of the user who manages the AI Dataset Asset record.

Table: User [user]

Data type: String
result.name Name of the associated AI Dataset Product Model record.

Table: AI Dataset Product Model [cmdb_ai_dataset_product_model]

Data type: String
result.provider Provider of the associated AI Dataset Product Model [cmdb_ai_dataset_product_model] table record.

Data type: Object

provider: {
  "name": "String",
  "sys_id": "String"
}
result.provider.sys_id Sys_id of the provider in the associated AI Dataset Product Model [cmdb_ai_dataset_product_model] table record.

Data type: String
result.provider.name Name of the provider.

Data type: String
result.source Details about the source of AI dataset asset.
Valid values:
  • Link to the source of the dataset asset.
  • Details (in plain text) of the source of the dataset asset. For example, the name of a product or website.

Data type: String

Default: empty or null
result.state State of the AI Dataset Asset record.
Possible values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String
result.sys_id Sys_id of the AI Dataset Asset record.

Table: AI Dataset Digital Asset [alm_ai_dataset_digital_asset]

Data type: String
result.updated Date and time that the AI Dataset Asset record was last updated.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.version Version number of the associated AI Dataset Product Model record. For example, V2.

Data type: String
result.warnings Comma-separated list of warning messages that are present when creating the dataset. These warnings can be validation checks, such as when the sys_id of an optional parameter is invalid.

Data type: Array

"warnings": ["String"]

Example: cURL request

The following example shows how to create a new AI dataset asset record.

curl -X POST 'https://instance.servicenow.com/api/sn_ent/asset/ai_dataset' \ 
  -H 'Accept: application/json' \ 
  -H 'Content-Type: application/json' \ 
  -u 'username':'password' \ 
  -d ' { 
  "name": "Dataset One", 
  "description": "Description for dataset ", 
  "provider": "servicenow", 
  "version": "V1", 
  "state": 31, 
  “source”: “Source of dataset” 
  "documentation": "document", 
  “dataset_card”: “Dataset Card”, 
  “base_datasets”: [ “Dataset Two”, “Dataset Three”], 
  “data_type”: “1,2”, 
  “acceptable_usage”: “1,2”, 
  "managed_by": "abel.tuter" 
}'

Response body. The results of the newly created dataset asset record.

{ 
  "result": { 
    "asset": { 
      "sys_id": "da8393eb40d25210f877b00c113d1fc1", 
      "display_name": "ServiceNow Closed Incidents", 
      "name": "Closed Incidents", 
      "description": "Incidents with resolution", 
      "documentation": "Sample Documentation", 
      "source": "incident table on servicenow instance", 
      "dataset_card": "Dataset Card", 
      "state": "Deployed", 
      "version": null, 
      "data_type": { 
        "value": "1", 
        "label": "Text" 
      }, 
      "provider": { 
        "sys_id": "93d4ecfac0a8000b6294d71b733977fb", 
        "name": "ServiceNow" 
      }, 
      "managed_by": { 
        "sys_id": "undefined", 
        "name": "" 
      }, 
      "acceptable_usage": { 
        "value": "1", 
        "label": "Training" 
      }, 
      "base_datasets": [], 
      "created": "2024-12-12 01:23:03", 
      "updated": "2024-12-12 01:23:03" 
    }, 
    "warnings": [] 
       } 
}

AI Assets API - POST /sn_ent/asset/ai_prompt

Creates a new AI prompt asset entry in the AI Prompt Digital Asset [alm_ai_prompt_digital_asset] and AI Prompt Product Model [cmdb_ai_promt_product_model] tables according to details you provide in the request body.

URL format

Versioned URL: /api/sn_ent/{api_version}/asset/ai_prompt

Default URL: /api/sn_ent/asset/ai_prompt

Supported request parameters

Table 31. Path parameters
Name Description
api_version Optional. Version of the endpoint to access. For example, v1 or v2. Only specify this value to use an endpoint version other than the latest.

Data type: String
Table 32. Query parameters
Name Description
None
Table 33. Request body parameters (XML or JSON)
Name Description
object Required. Content of the AI Prompt asset to update.

Data type: Object

{
  "ai_model": "String",
  "description": "String",
  "documentation": "String",
  "managed_by": "String" 
  "name": "String",
  "prompt_info": "String", 
  "provider": "String",
  "state": Number,
  "version": "String"
}
ai_model Value of the AI Model field of an existing record in the AI Model Digital Asset [alm_ai_model_digital_asset] table.
Valid values:
  • Name of the AI Model Digital Asset [alm_ai_model_digital_asset] record
  • Sys_id of the AI Model Digital Asset [alm_ai_model_digital_asset] record

Data type: String

Default: empty string
description Description to give the AI Prompt Product Model.

Table: AI Prompt Product Model [cmdb_ai_prompt_product_model]

Data type: String

Default: empty string
documentation Documentation of the AI Prompt Product Model.

Data type: String

Default: empty string
managed_by Value of the Managed by field of an existing record in the User [sys_user] table.
Valid values:
  • Name of the User [sys_user] record
  • Sys_id of the User [sys_user] record

Data type: String

Default: empty string
name Required. Name of the associated record in the AI Prompt Product Model [cmdb_ai_prompt_product_model] table.

Data type: String

Default: empty string
prompt_info Prompt information for AI Prompt Asset.

Data type: String

Default: empty string
provider Required. Value of the Provider field of an existing record in the Company [core_company] table.
Valid values:
  • Name of the Company [core_company] record
  • Sys_id of the Company [core_company] record

Data type: String
state State to apply to the AI Model Asset. For example,
Valid values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String
version Version number to assign the AI Prompt Product Model. For example, V2.

Data type: String

Default: empty string

Headers

The following request and response headers apply to this HTTP action only, or apply to this action in a distinct way. For a list of general headers used in the REST API, see Supported REST API headers.

Table 34. Request headers
Header Description
Accept Data format of the response body. Only supports application/json.
Content-Type Data format of the request body. Only supports application/json.
Table 35. Response headers
Header Description
None

Status codes

The following status codes apply to this HTTP action. For a list of possible status codes used in the REST API, see REST API HTTP response codes.

Table 36. Status codes
Status code Description
200 Successful. The request was successfully processed.
400 Bad Request. A bad request type or malformed request was detected.
500 Internal server error. An unexpected error occurred while processing the request. The response contains additional information about the error.

Response body parameters (JSON or XML)

Name Description
result Results of the new AI asset prompt.

Data type: Object

"result": { 
    "asset": Object, 
    "warnings": [Array] 
 }
result.asset Details about the newly created entry.

Data type: Object

"asset": {
  "ai_model": [Array],
  "ai_prompts": [Array]
  "created": "String",
  "description": "String",
  "display_name": "String",
  "documentation": "String",
  "evaluation_datasets": [Array],
  "evaluation_metrics_report": String,
  "managed_by": Object,
  "name": "String",
  "provider": Object,
  "state": "String",
  "sys_id": "String",
  "updated": "String",
  "version": "String",
  "warnings": [Array]
}
result.asset.ai_model List of AI models in the AI System Asset [cmdb_ai_ system_asset_model] table record.

Data type: Array

"ai_model": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.asset.ai_models.name Name of the AI System Digital Asset [alm_ai_system_digital_asset] table record.

Data type: String
result.asset.ai_models.sys_id Sys_id of the AI System Digital Asset [alm_ai_system_digital_asset] table record.

Data type: String
result.asset.ai_prompts List of AI Prompts of the AI System Asset record.

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: Array

"ai_prompts": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.asset.ai_prompts.name Name of the AI prompt.

Data type: String
result.asset.ai_prompts.sys_id Sys_id of the AI System Digital Asset [alm_ai_system_digital_asset] table record.

Data type: String
result.asset.created Date and time that the AI model asset was created.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.asset.description Description of the associated AI System Product Model record.

Table: AI System Product Model [cmdb_ai_system_product_model]

Data type: String
result.asset.display_name Display name of the AI System Asset record.

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: String
result.asset.documentation Documentation of the associated AI System Product Model [cmdb_ai_ system_product_model] table record.

Data type: String
result.asset.evaluation_datasets List of sys_ids or display names of the AI datasets used to evaluate the AI system asset model. Mostly applicable for models developed within an organization.

Tables: AI Dataset Digital Asset [alm_ai_dataset_digital_asset], AI System Digital Asset [alm_ai_system_digital_asset]

Data type: Array

"evaluation_datasets": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.asset.evaluation_datasets.name Name of the AI Dataset Digital Asset.

Data type: String
result.asset.evaluation_datasets.sys_id Sys_id of the AI System Digital Asset record.

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: String
result.asset.evaluation_metrics_report Evaluation results of the AI system asset.
Possible values:
  • Details (in plain text) outlining results
  • Links to specific results

Data type: String
result.asset.managed_by Details about the user who manages the asset.

Data type: Object

"managed_by": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.asset.managed_by.name Name of the user who manages the AI model asset record.

Table: User [user]

Data type: String
result.asset.managed_by.sys_id Sys_id of the user who manages the AI model asset record.

Table: User [user]

Data type: String
result.asset.name Name of the associated AI System Product Model record.

Table: AI System Product Model [cmdb_ai_system_product_model]

Data type: String
result.asset.provider Value of the Provider field in the associated AI System Product Model [cmdb_ai_ system_product_model] table record.

Data type: Object

provider: {
  "name": "String",
  "sys_id": "String"
}
result.asset.provider.name Name of the provider in the associated AI System Product Model [cmdb_ai_ system_product_model] table record.

Data type: String
result.asset.provider.sys_id Sys_id of the provider in the associated AI System Product Model [cmdb_ai_ system_product_model] table record.

Data type: String
result.asset.state State of the AI Model Asset.
Possible values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String
result.asset.sys_id Sys_id of the AI System Asset record.

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: String
result.asset.updated Date and time that the AI model asset was last updated.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.warning Comma-separated list of warning messages. These warnings can be validation checks, such as when the sys_id of an optional parameter is invalid.

Data type: Array

"warnings": ["String"]

Example: cURL request

The following example shows how to use the POST method to create a new AI prompt according to details provided in the request body.

curl -X POST 'https://instance.servicenow.com/api/sn_ent/asset/ai_prompt' \ 
  -H 'Accept: application/json' \ 
  -H 'Content-Type: application/json' \ 
  -u 'username':'password' \ 
  -d '{ 
  "name": "Incident Summarization prompt1", 
  "description": "Prompt for Incident Summarization", 
  "provider": "servicenow", 
  "version": "V1", 
  "state": 31, 
  "documentation": "Document", 
  "ai_model": "mixtral-instruct", 
  "prompt_info": "Provide incident summary using short_decription, state, worknotes", 
  "managed_by": "abel.tuter" 
 }'

Response body.

{ 
  "result": { 
    "asset": { 
      "sys_id": "9833721b331e92101c9aca989d5c7bf0", 
      "display_name": "ServiceNow Incident Summarization prompt1 V1", 
      "name": "Incident Summarization prompt1", 
      "description": "Prompt for Incident Summarization", 
      "version": "V1", 
      "provider": { 
        "sys_id": "93d4ecfac0a8000b6294d71b733977fb", 
        "name": "ServiceNow" 
      }, 
      "documentation": "Document", 
      "state": "Deployed", 
      "ai_model": { 
        "sys_id": "9d7dc7e6eb1e5210aa82fab8bad0cda2", 
        "name": "mixtral-instruct" 
      }, 
      "prompt_info": "Provide incident summary using short_decription, state, worknotes", 
      "managed_by": { 
        "sys_id": "62826bf03710200044e0bfc8bcbe5df1", 
        "name": "Abel Tuter" 
      }, 
      "created": "2024-12-11 04:23:17", 
      "updated": "2024-12-11 04:23:17" 
    }, 
    "warnings": [] 
  } 
}

AI Assets API - POST /sn_ent/asset/ai_model

Creates a new AI model asset entry in the AI Model Digital Asset [alm_ai_model_digital_asset] and AI Model Product Model [cmdb_ai_model_product_model] tables according to details you provide in the request body.

URL format

Versioned URL: /api/sn_ent/{api_version}/asset/ai_model

Default URL: /api/sn_ent/asset/ai_model

Supported request parameters

Table 37. Path parameters
Name Description
api_version Optional. Version of the endpoint to access. For example, v1 or v2. Only specify this value to use an endpoint version other than the latest.

Data type: String
Table 38. Query parameters
Name Description
None
Table 39. Request body parameters (XML or JSON)
Name Description
{object} Required. Details to apply to the new asset model record.

Data type: Object

"object": {
  "base_model": {Object},
  "context_window:" "String",
  "deployment_guideline": "String",
  "description": "String", 
  "documentation": "String",
  "evaluation_datasets": [Array],
  "evaluation_metrics_report": "String",
  "managed_by": "String", 
  "model_size_in_mb": "String",
  "model_weights_info": "String",
  "name": "String",
  "parameters_info": "String",
  "provider": "String",
  "required_infrastructure": "String",
  "source": "String",
  "state": Number,
  "supported_languages: [Array],
  "training_datasets": [Array],
  "training_procedure": "String",
  "version": "String"
}
{object}.base_model AI model that this model version was derived from.
Note: Only applicable for models developed within the organization.

Data type: Object

{
  "name": "String",
  "sys_id": "String"
 }
{object}.base_model.name Name of the AI model asset to model this AI model after.

Data type: String
{object}.base_model.sys_id Sys_id of the AI model asset to model this AI model after.

Table: AI Model Digital Asset [alm_ai_dataset_digital_asset]

Data type: String

Default: empty string
{object}.context_window Size of input sequences (in other words, the number of tokens) that the model can handle.

Data type: Integer

Default: empty value
{object}.deployment_guideline Instructions applicable for models developed and deployed within an organization.

Data type: String

Default: empty string
{object}.description Description to give the AI Model Product Model.

Updated table: AI Model Product Model [cmdb_ai_model_product_model]

Data type: String

Default: empty string
{object}.documentation Documentation of the AI Model Product Model.

Data type: String

Default: empty string
{object}.evaluation_datasets Comma-separated list of sys_ids or display names of AI datasets of the AI datasets used for evaluating the model. Mostly applicable for models developed within an organization.

Tables: AI Dataset Digital Asset [alm_ai_dataset_digital_asset], AI Model Asset [alm_ai_model_digital_asset]

Data type: Array

"evaluation_datasets": ["String", "String"]
Valid values:
  • Name of the AI Dataset Digital Asset record.
  • Sys_id of the AI Dataset Digital Asset record.

Default: empty string
{object}.evaluation_metrics_report Reference to the evaluation results located within the text field of the AI model digital asset record. For example:
Testing results:  link to the result document

Details:
Accuracy: 85%
Hallucination: 10%
Eval Run 1: link
Eval Run 2: link

Table: AI model digital asset [alm_ai_model_digital_asset]

Data type: String

Default: empty string
{object}.managed_by Value of the 'Managed by' field of an existing record in the User [sys_user] table.
Valid values:
  • Name of the User [sys_user] record
  • Sys_id of the User [sys_user] record

Data type: String

Default: empty string
{object}.model_size_in_mb Size of the model in MB. Mostly applicable for models developed and deployed within an organization.

Data type: Number

Default: null or empty
{object}.model_weights_info Additional model information, if available. Mostly applicable for models developed within an organization.

Data type: String

Default: null or empty
{object}.name Required. Name of the AI Model Product Model.

Table: AI Model Product Model [cmdb_ai_model_product_model]

Data type: String

Default: null or empty
{object}.parameters_info Number of parameters given to for the model.

Data type: String

Default: empty string
{object}.provider Required. Value of the 'Provider' field of existing record in Company [core_company] table. Assigns this provider to the new AI model.
Valid values:
  • Name of the Company [core_company] record
  • Sys_id of the Company [core_company] record

Data type: String

Default: empty string
{object}.required_infrastructure Documentation of infrastructure needs for the model deployment. For example, details about the infrastructure stack and processing needs. Mostly applicable for models deployed within an organization.

Data type: String

Default: empty or null
{object}.source Details about the source of who or what created the model.
Valid values:
  • Link to the source of the model. For example, https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1 or a link to Github.
  • Details of the source in plain text. For example, Microsoft Azure

Data type: String

Default: empty or null
{object}.state State to apply to the AI Model Asset.
Valid values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String
{object}.supported_languages Comma-separated list of languages that are supported by the AI Model.

Data type: Array

"supported_languages": [
  "String",
  "String"
]
Valid values:
  • Sys_id of Language [sys_language] record
  • Name of Language [sys_language] record. For example, "French", "English".

Default: empty string
{object}.training_datasets Datasets used to train the AI Model. Comma-separated list of sys_ids or display names of the AI Dataset Asset [alm_ai_dataset_digital_asset] table. Mostly applicable for models developed within an organization.

Data type: Array

"training_datasets": [
  "String",
  "String"
]

Default: empty string
{object}.training_procedure Type of training to apply to the AI model.
Valid values:
  • 1: Decision Trees
  • 2: Deep Neural Networks
  • 3: Linear Regression
  • 4: Logistic Regression
  • 5: Random Forest
  • 6: Supervised Learning
  • 7: Unsupervised Learning
  • 8: Reinforcement Learning
  • 9: Transfer Learning
  • 10: Semi-Supervised Learning
  • 11: Instruction Finetuning
  • 12: Supervised Finetuning

Data type: String represented with a number (for example, "3")

Default: 1: Decision Trees
{object}.version Version for AI Model Product Model. For example, V2.

Data type: String

Default: empty string

Response body parameters (JSON or XML)

Headers

The following request and response headers apply to this HTTP action only, or apply to this action in a distinct way. For a list of general headers used in the REST API, see Supported REST API headers.

Table 40. Request headers
Header Description
Accept Data format of the response body. Only supports application/json.
Content-Type Data format of the request body. Only supports application/json.
Table 41. Response headers
Header Description
None

Status codes

The following status codes apply to this HTTP action. For a list of possible status codes used in the REST API, see REST API HTTP response codes.

Table 42. Status codes
Status code Description
201 Successful. The request was successfully processed.
400 Bad Request. A bad request type or malformed request was detected.
500 Internal server error. An unexpected error occurred while processing the request. The response contains additional information about the error.

Response body parameters (JSON or XML)

Name Description
result Results of the AI model asset creation.

Data type: Object

"result": { 
    "asset": Object, 
    "warnings": [Array], 
 }
result.asset Details about the created AI model asset.

Data type: Object

"asset": {
  "base_model": Object,
  "context_window": String,
  "created": "String",
  "deployment_guideline": String,
  "description": "String",
  "display_name": "String",
  "documentation": "String",
  "evaluation_datasets": Array,
  "evaluation_metrics_report": String,
  "managed_by": Object,
  "model_size_in_mb": String,
  "name": "String",
  "parameters_info": "String",
  "provider": Object,
  "required_infrastructure": String,
  "state": "String",
  "source": String, 
  "supported_languages": Array,
  "sys_id": "String",
  "training_datasets": Array,
  "training_procedure": String,
  "updated": "String",
  "version": "String"
}
result.asset.base_model Information about the AI model asset from which this version derived from.

Data type: Object

"base_model": { 
  "name": "String",
  "sys_id": "String"
}
result.asset.base_model.name Display name of AI model asset.

Table: AI Model Digital Asset [alm_ai_model_digital_asset]

Data type: String
result.asset.base_model.sys_id Sys_id of AI model asset.

Table: AI Model Digital Asset [alm_ai_model_digital_asset]

Data type: String
result.asset.context_window Size of input sequences that the model can handle. In other words, the number of tokens.

Data type: String represented with a number. For example, "6000".
result.asset.created Date and time that the AI model asset was created.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.asset.deployment_guideline Instructions applicable for models developed and deployed within an organization.

Data type: String
result.asset.description Description of the associated AI Model Product Model record.

Table: AI Model Product Model [cmdb_ai_model_product_model]

Data type: String
result.asset.display_name Display name of the AI Model Asset record.

Table: AI Model Digital Asset [alm_ai_model_digital_asset] (display_name field)

Data type: String
result.asset.documentation Documentation of the associated AI Model Product Model record.

Table: AI Model Product Model [cmdb_ai_model_product_model]

Data type: String
result.asset.evaluation_datasets Comma-separated list of sys_ids or display names of AI datasets of the AI Model Digital Asset used for evaluating the model. Mostly applicable for models developed within an organization.

Tables: AI Dataset Digital Asset [alm_ai_dataset_digital_asset], AI Model Digital Asset [alm_ai_model_digital_asset]

Data type: Array

"evaluation_datasets": [
  { 
  "name": "String",
  "sys_id": "String"
  } 
]
result.evaluation_datasets.name Name of the AI Dataset Digital Asset.

Data type: String
result.evaluation_datasets.sys_id Sys_id of the AI Model Digital Asset record.

Table: AI Model Digital Asset [alm_ai_model_digital_asset]

Data type: String
result.asset.evaluation_metrics_report Reference to the evaluation results.
Possible values:
  • Details (in plain text) outlining results
  • Links to specific results

Data type: String
result.asset.managed_by User that manages the AI model asset record.

Data type: Object

" managed_by": {
  "name": "String",
  "sys_id": "String"
}
result.asset.managed_by.name Name of the user who manages the AI model asset record.

Table: User [user]

Data type: String
result.asset.managed_by.sys_id Sys_id of the user who manages the AI model asset record.

Table: User [user]

Data type: String
result.asset.model_size_in_mb Size of the model in MB. Usually applicable for models developed and deployed within an organization.

Data type: Number
result.asset.name Name of the associated AI Model Product Model record.

Table: AI Model Product Model [cmdb_ai_model_product_model]

Data type: String
result.asset.provider Provider of the associated AI Model Product Model record.

Table: AI Model Product Model [cmdb_ai_model_product_model]

Data type: Object

"provider": {
  "name": "String",
  "sys_id": "String"
}
result.asset.provider.name Name of the provider.

Data type: String
result.asset.provider.sys_id Sys_id of the record from Company [core_company] table that corresponds to the Provider of the associated AI Model Product Model record.

Table: AI Model Product Model [cmdb_ai_model_product_model]

Data type: String
result.asset.source Details about the source of the asset.
Valid values:
  • Link to the source of the model. For example, https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1 or a link to Github.
  • Details of the source in plain text. For example, Microsoft Azure

Data type: String
result.asset.state State of the AI Model Asset record.
Possible values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String
result.asset.supported_languages.name Name of the supported language.

Table: Language [sys_language]

Data type: String
result.asset.supported_languages.sys_id Sys_id of the supported language.

Table: Language [sys_language]

Data type: String
result.asset.sys_id Sys_id of the AI Model Asset record

Table: AI Model Digital Asset [alm_ai_model_digital_asset]

Data type: String
result.asset.training_datasets Reference to 1+ associated data sets used for training the model.

Data type: Array

"training_datasets": [
  { 
  "name": "String",
  "sys_id": "String"
  } 
]
result.asset.training_procedure Type of AI training applied to the model.
Possible values:
  • 1: Decision Trees
  • 2: Deep Neural Networks
  • 3: Linear Regression
  • 4: Logistic Regression
  • 5: Random Forest
  • 6: Supervised Learning
  • 7: Unsupervised Learning
  • 8: Reinforcement Learning
  • 9: Transfer Learning
  • 10: Semi-Supervised Learning
  • 11: Instruction Finetuning
  • 12: Supervised Finetuning

Data type: String
result.asset.updated Date and time that the AI model asset was last updated.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.asset.version Version of the associated AI Model Product Model record.

Table: AI Model Product Model [cmdb_ai_model_product_model table]

Data type: String
result.required_infrastructure Documentation of infrastructure needs for the model's deployment. For example, details about the infrastructure stack and processing needs.

Data type: String
result.asset.parameters_info Properties of the training data that learn during the learning process. For example: 7B or 30B.

Data type: String
result.asset.supported_languages Details about the languages supported by the AI model asset.

Table: Language [sys_language]

Data type: Object

"supported_languages": [
 { 
  "name": "String",
  "sys_id": "String"
  }
]
result.warnings Comma-separated list of warning messages. These warnings can be validation checks, such as when the sysId of an optional parameter is invalid.

Data type: Array

"warnings": ["String"]

Example: cURL request

The following example creates an AI model asset according to details provided in the request body.

curl -X POST 'https://instance.servicenow.com/api/sn_ent/asset/ai_prompt' \ 
  -H 'Accept: application/json' \ 
  -H 'Content-Type: application/json' \ 
  -u 'username':'password' \ 
  -d ' {
  "name": "Now LLM",
  "description": "enables text-to-text like question answering and summarization",
  "provider": "servicenow",
  "documentation": "Now LLM V5 Documentation",
  "version": "V8",
  "parameters_info": "7B",
  "supported_languages": [
    "English",
    "French"
  ],
  "model_size_in_mb": "87",
  "deployment_guideline": "Deployed on ServiceNow infrastructure",
  "source": "https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1",
  "training_procedure": "2",
  "context_window": "8000",
  "state": "31",
  "base_model": "Servicenow Now LLM V6",
  "model_weights_info": "refer to weights and biases project",
  "required_infrastructre": "GPUs needed: 1, GPU Type: A100",
  "training_datasets": [
    "Servicenow Open Incidents"
  ],
  "evaluation_datasets": [
    "Servicenow Open Incidents"
  ],
  "evaluation_metrics_report": "Testing results: See files attached to this record",
  "managed_by": "abel.tuter"
}'
The response body shows details about the newly created AI model asset, including the resultant sys_id.
{
  "result": {
    "asset": {
      "sys_id": "a438d170ff96da10c1fbffffffffffd5",
      "display_name": "ServiceNow Now LLM V6",
      "name": "Now LLM",
      "description": "enables text-to-text like question answering and summarization",
      "version": "V8",
      "provider": {
        "sys_id": "93d4ecfac0a8000b6294d71b733977fb",
        "name": "ServiceNow"
      },
      "documentation": "Now LLM V5 Documentation",
      "parameters_info": "7B",
      "supported_languages": [
        {
          "sys_id": "914493a30f320010e96b0e4fef767e90",
          "name": "English"
        }
      ],
      "model_size_in_mb": "87",
      "deployment_guideline": "Deployed on ServiceNow infrastructure",
      "source": "https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1",
      "training_procedure": "2",
      "context_window": "8000",
      "state": "Deployed",
      "required_infrastructure": "Servicenow Instance",
      "base_model": {
        "sys_id": "a438d170ff96da10c1fbffffffffffd5",
        "name": "ServiceNow Now LLM V8"
      },
      "evaluation_datasets": [
        {
          "sys_id": "45cb45baff06d610c1fbffffffffffa9",
          "name": "ServiceNow Open Incidents"
        }
      ],
      "training_datasets": [
        {
          "sys_id": "45cb45baff06d610c1fbffffffffffa9",
          "name": "ServiceNow Open Incidents"
        }
      ],
      "evaluation_metrics_report": "Testing results: See files attached to this record",
      "managed_by": {
        "sys_id": "62826bf03710200044e0bfc8bcbe5df1",
        "name": "Abel Tuter"
      },
      "created": "2024-12-03 16:50:53",
      "updated": "2024-12-12 15:56:28"
    },
    "warnings": [
      "Reference record 'French' not found for supported_languages in table sys_language"
    ]
  }
}

AI Assets API - POST /sn_ent/asset/ai_system

Creates a new AI system asset entry in the AI System Digital Asset [alm_ai_system_digital_asset] and AI System Product Model [cmdb_ai_system_product_model] tables according to details you provide in the request body.

URL format

Versioned URL: /api/sn_ent/{api_version}/asset/ai_system

Default URL: /api/sn_ent/asset/ai_system

Supported request parameters

Table 43. Path parameters
Name Description
api_version Optional. Version of the endpoint to access. For example, v1 or v2. Only specify this value to use an endpoint version other than the latest.

Data type: String
Table 44. Query parameters
Name Description
None
Table 45. Request body parameters (XML or JSON)
Name Description
object Required. Details to update in the AI system.

Data type: Object

{ 
  "name": "String", 
  "description": "String", 
  "provider": "String", 
  "version": "String", 
  "state": Number, 
  "documentation": "String", 
  "ai_models": "String", 
  "ai_prompts": "Strings", 
  "evaluation_datasets": [Array], 
  "evaluation_metrics_report": "String", 
  "managed_by": "String" 
}
object.name Required. Name for the AI System Product Model.

Data type: String
object.description Description for the AI System Product Model.

Data type: String
object.provider Required. Value of the Provider field in an existing Company [core_company] table record.
Valid values:
  • Name of the Company [core_company] record
  • Sys_id of the Company [core_company] record

Data type: String
object.state State to apply to the AI System Asset.
Valid values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String

Default: empty string
object.version Version for AI System Product Model. For example, V2.

Data type: String

Default: empty string
object.documentation Documentation of the AI System Asset.

Data type: String

Default:
object.ai_models Comma-separated list of strings, where each string represents an AI Model field value of an existing AI Model Digital Asset record.

Table: AI Model Digital Asset [alm_ai_model_digital_asset]

Valid values:
  • Name of the AI Model Digital Asset record.
  • Sys_id of the AI Model Digital Asset record.

Data type: String

Default: empty string
object.evaluation_datasets Comma-separated list of AI datasets used for evaluating the AI System Asset record. Mostly applicable for models developed within an organization.

Tables: AI Dataset Digital Asset [alm_ai_dataset_digital_asset], AI System Asset [AI System Digital Asset alm_ai_system_digital_asset]

Data type: Array

"evaluation_datasets": [{String", "String"}]
Valid values:
  • Name of the AI Dataset Digital Asset record.
  • Sys_id of the AI Dataset Digital Asset record.

Default: empty string
object.evaluation_metrics_report Reference to the evaluation results located within the text field of the AI system asset record. For example:
Testing results:  link to the result document

Details:
Accuracy: 85%
Hallucination: 10%
Eval Run 1: link
Eval Run 2: link

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: String

Default: empty string
object.managed_by Value of the 'Managed by' field of an existing User [sys_user] table record.
Valid values:
  • Name of the User [sys_user] record
  • Sys_id of the User [sys_user] record

Data type: String

Default: empty string

Headers

The following request and response headers apply to this HTTP action only, or apply to this action in a distinct way. For a list of general headers used in the REST API, see Supported REST API headers.

Table 46. Request headers
Header Description
Accept Data format of the response body. Only supports application/json.
Content-Type Data format of the request body. Only supports application/json.
Table 47. Response headers
Header Description
None

Status codes

The following status codes apply to this HTTP action. For a list of possible status codes used in the REST API, see REST API HTTP response codes.

Table 48. Status codes
Status code Description
200 Successful. The request was successfully processed.
400 Bad Request. A bad request type or malformed request was detected.
500 Internal server error. An unexpected error occurred while processing the request. The response contains additional information about the error.

Response body parameters (JSON or XML)

Name Description
result Results of the new AI system asset.

Data type: Object

"result": {
  "asset": Object,
  "warnings": [Array]
}
result.asset Details about the newly created entry.

Data type: Object

"asset": {
  "ai_model": [Array],
  "ai_prompts": [Array]
  "created": "String",
  "description": "String",
  "display_name": "String",
  "documentation": "String",
  "evaluation_datasets": [Array],
  "evaluation_metrics_report": String,
  "managed_by": Object,
  "name": "String",
  "provider": Object,
  "state": "String",
  "sys_id": "String",
  "updated": "String",
  "version": "String",
  "warnings": [Array]
}
result.asset.ai_models List of AI models in the AI System Digital Asset [alm_ai_system_digital_asset] record.

Data type: Array

"ai_models": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.asset.ai_models.name Name of the AI System Digital Asset [alm_ai_system_digital_asset] table record.

Data type: String
result.asset.ai_models.sys_id Sys_id of the AI System Digital Asset [alm_ai_system_digital_asset] table record.

Data type: String
result.asset.ai_prompts List of AI Prompts of the AI System Digital Asset [alm_ai_system_digital_asset] record.

Data type: Array

"ai_prompts": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.asset.ai_prompts.name Name of the AI prompt.

Data type: String
result.asset.ai_prompts.sys_id Sys_id of the AI System Digital Asset [alm_ai_system_digital_asset] table record.

Data type: String
result.asset.created Date and time that the AI model asset was created.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.asset.description Description of the associated AI System Product Model record.

Table: AI System Product Model [cmdb_ai_system_product_model]

Data type: String
result.asset.display_name Display name of the AI System Asset record.

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: String
result.asset.documentation Documentation of the associated AI System Product Model [cmdb_ai_ system_product_model] table record.

Data type: String
result.asset.evaluation_datasets List of AI datasets used for evaluating the AI System Asset record. Mostly applicable for models developed within an organization.

Tables: AI Dataset Digital Asset [alm_ai_dataset_digital_asset], AI System Asset [alm_ai_system_digital_asset]

Data type: Array

"evaluation_datasets": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.asset.evaluation_datasets.name Name of the AI Dataset Digital Asset.

Data type: String
result.asset.evaluation_datasets.sys_id Sys_id of the AI Dataset Digital Asset [alm_ai_dataset_digital_asset] table record.

Data type: String
result.asset.evaluation_metrics_report Evaluation results of the AI system asset.
Possible values:
  • Details (in plain text) outlining results
  • Links to specific results

Data type: String
result.asset.managed_by Comma-separated list of sys_ids or display names of AI datasets of the AI System Asset used for evaluating the model. Mostly applicable for models developed within an organization.

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: Object

"managed_by": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.asset.managed_by.name Sys_id of the user who manages the AI model asset record.

Table: User [user]

Data type: String
result.asset.managed_by.sys_id Name of the user who manages the AI model asset record.

Table: User [user]

Data type: String
result.asset.name Name of the associated AI System Product Model record.

Table: AI System Product Model [cmdb_ai_system_product_model]

Data type: String
result.asset.provider Value of the Provider field in the associated AI System Product Model [cmdb_ai_ system_product_model] table record.

Data type: Object

provider: {
  "name": "String",
  "sys_id": "String"
}
result.asset.provider.name Name of the provider in the associated AI System Product Model [cmdb_ai_ system_product_model] table record.

Data type: String
result.asset.provider.sys_id Sys_id of the provider in the associated AI System Product Model [cmdb_ai_ system_product_model] table record.

Data type: String
result.asset.state State of the AI Model Asset.
Possible values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String
result.asset.sys_id Sys_id of the AI System Asset record.

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: String
result.asset.updated Date and time that the AI model asset was last updated.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.asset.version Version number of the associated AI System Product Model record. For example, V2.

Data type: String
result.asset.warnings Comma-separated list of warning messages. These warnings can be validation checks, such as when the sys_id of an optional parameter is invalid.

Data type: Array

"warnings": ["String"]

Example: cURL request

The following example inserts a new entry into the AI System Digital Asset and AI System Product Model tables using details included in the request body.

curl -X POST 'https://instance.servicenow.com/api/sn_ent/asset/ai_prompt' \ 
  -H 'Accept: application/json' \ 
  -H 'Content-Type: application/json' \ 
  -u 'username':'password' \ 
  -d '{ 
  "name": "Incident Summarization", 
  "description": "Incident Summarization Skill", 
  "provider": "servicenow", 
  "documentation": "Sample Documentation", 
  "version": "V2", 
  "state": 31, 
  "ai_models": [ 
    "llm_generic_small", 
    "mixtral-instruct" 
  ], 
  "ai_prompts": [ 
    "LLM Prompt"   
  ], 
  "evaluation_datasets": [ 
    "Base dataset" 
  ], 
  "evaluation_metrics_report": "Sample Report", 
  "managed_by": "abel.tuter" 
}'

Response body.

{ 
  "result": { 
    "asset": { 
      "sys_id": "3b140397435a9210a63d00002fb8f2d7", 
      "display_name": "ServiceNow Incident Summarization V2", 
      "name": "Incident Summarization", 
      "description": "Incident Summarization Skill", 
      "version": "V2", 
      "provider": { 
        "sys_id": "93d4ecfac0a8000b6294d71b733977fb", 
        "name": "ServiceNow" 
      }, 
      "documentation": "Sample Documentation", 
      "state": "Deployed", 
      "ai_models": [{  
        "sys_id": "9tgdc7e6eb1e5210aa82fab8bad0cda2",  
        "name": "llm_generic_small"  
      }, 
      {  
        "sys_id": "7efdc7e6eb1e5210aa82fab8bad0cda2",  
        "name": "mixtral-instruct"  
      }], 
      "ai_prompts": [{  
        "sys_id": "7d7dc7e6eb1e5210aa82fab8bad0cda2",  
        "name": "LLM Prompt"  
      }], 
      "evaluation_datasets": [{  
        "sys_id": "9d7dc7e6eb1e5210aa82fab8bad0cda2",  
        "name": "Base dataset"  
      }], 
      "evaluation_metrics_report": "Sample Report", 
      "managed_by": { 
        "sys_id": "62826bf03710200044e0bfc8bcbe5df1", 
        "name": "Abel Tuter" 
      }, 
      "created": "2024-12-11 18:23:09", 
      "updated": "2024-12-11 18:23:09" 
    }, 
    "warnings": [] 
  } 
}

AI Assets API - PUT /sn_ent/asset/ai_dataset/{sys_id}

Updates the data of a specific AI dataset asset record according to information that you provide in the request body.

Note: Provide only the parameter-value pairs for the specific data you want to update. This endpoint overwrites the data for any parameters that are sent in the request.

Use the AI Assets API - GET /sn_ent/asset/ai_dataset/{sys_id} method to retrieve an existing AI dataset record with a given ID. You can then use this PUT method to update values in the dataset using the same ID information.

URL format

Versioned URL: /api/sn_ent/{api_version}/asset/ai_dataset/{sys_id}

Default URL: /api/sn_ent/asset/ai_dataset/{sys_id}

Supported request parameters

Table 49. Path parameters
Name Description
api_version Optional. Version of the endpoint to access. For example, v1 or v2. Only specify this value to use an endpoint version other than the latest.

Data type: String
sys_id Sys_id of the asset in the AI Dataset Asset [alm_ai_dataset_digital_asset] table.

Data type: String
Table 50. Query parameters
Name Description
None
Table 51. Request body parameters (XML or JSON)
Name Description
{object} 
{
  "acceptable_usage": "String",
  "base_datasets": [Array],
  "dataset_card": "String"
  "data_type": "String",
  "description": "String",
  "documentation": "String",
  "managed_by": "String", 
  "name": "String", 
  "provider": "String",
  "state": "String" 
  "source": "String",
  "version": "String"
}
{object}.acceptable_usage Determines how a dataset or model can be used, typically for training or evaluation purposes.
Valid values:
  • 1: Training
  • 2: Evaluation

Data type: String
{object}.base_datasets Comma-separated list of base datasets needed to build this dataset. Accepts names or sys_ids of datasets present in AI Dataset Digital Asset [alm_ai_dataset_digital_asset] table.

Data type: Array

“base_datasets”: [ “String”, “String”]
{object}.data_type Type of data present in the dataset. For example, Text,Video,Image or 1,2.

Data type: String
{object}.dataset_card The data set card. A dataset_card is a metadata document that describes the contents, structure, and context of an AI dataset. It provides details like data sources, features, intended use, and any known limitations to ensure proper understanding and usage.

Data type: String
{object}.description Description of the associated record in the AI Dataset Product Model [cmdb_ai_dataset_product_model] table.

Data type: String
{object}.documentation Documentation for the AI Dataset Product Model.

Data type: String
{object}.managed_by Value of the Managed by field of an existing User [sys_user] table record.
Valid values:
  • Name of the User [sys_user] record
  • Sys_id of the User [sys_user] record

Data type: String
{object}.name Required. Name of the associated record in the AI Dataset Product Model [cmdb_ai_dataset_product_model] table.

Data type: String
{object}.provider Required. Value of the Provider field of an existing record in the Company [core_company] table.
Valid values:
  • Name of the Company [core_company] record
  • Sys_id of the Company [core_company] record

Data type: String
{object}.source Details about the source of the dataset.
Valid values:
  • Link to the source of the dataset.
  • Details of the source in plain text.

Data type: String

Default: empty or null
{object}.state State of the AI dataset asset.
Valid values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String
{object}.version Version number of the associated AI Dataset Product Model record. For example, V2.

Data type: String

Headers

The following request and response headers apply to this HTTP action only, or apply to this action in a distinct way. For a list of general headers used in the REST API, see Supported REST API headers.

Table 52. Request headers
Header Description
Accept Data format of the response body. Only supports application/json.
Content-Type Data format of the request body. Only supports application/json.
Table 53. Response headers
Header Description
None

Status codes

The following status codes apply to this HTTP action. For a list of possible status codes used in the REST API, see REST API HTTP response codes.

Table 54. Status codes
Status code Description
200 Successful. The request was successfully processed.
400 Bad Request. A bad request type or malformed request was detected.
404 Not found. The requested item wasn't found.
500 Internal server error. An unexpected error occurred while processing the request. The response contains additional information about the error.

Response body parameters (JSON or XML)

Name Description
result Details of the newly created AI Dataset asset.

Data type: Object

"result": {
  "acceptable_usage": {Object},
  "base_datasets": [Array],
  "created": "String",
  "dataset_card": "String",
  "data_type": {Object},
  "description": "String",
  "documentation": "String",
  "display_name": "String",
  "managed_by": {Object},
  "name": "String",
  "provider": {Object},
  "state": "Development",
  "source": "String",
  "sys_id": "String",
  "updated": "String",
  "version": "String",
}
result.acceptable_usage Acceptable usage for the AI Dataset Asset record. Acceptable usage refers to how a dataset or model can be used, typically for training or evaluation purposes.

Data type: Object

"acceptable_usage": {
  "label": "String" 
  "value": "String" 
}
result.acceptable_usage.label Display label of the acceptable usage value.

Data type: String
result.acceptable_usage.value Number value of the acceptable usage.
Valid values:
  • 1: Training
  • 2: Evaluation

Data type: String
result.base_datasets Comma-separated list of base datasets required to build the given dataset. Accepts the name or sys_id of a base data set in the AI Dataset Digital Asset [alm_ai_dataset_digital_asset] table.

Data type: Array

"base_datasets": ["String", "String"]
result.created Date and time that the AI Dataset Asset record was created.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.data_type The type of data present in the AI Dataset Asset record.

Data type: Object

"data_type": {
  "label": "String",
  "value": "String"
}
result.data_type.label The display label for the data type value.

Data type: String
result.data_type.value Value of the dataset asset's data type.

Data type: String
result.dataset_card The data set card. A dataset_card is a metadata document that describes the contents, structure, and context of an AI dataset. It provides details like data sources, features, intended use, and any known limitations to ensure proper understanding and usage.

Data type: String
result.description Description of the associated AI Dataset Product Model record.

Table: AI Dataset Product Model [cmdb_ai_dataset_product_model]

Data type: String
result.display_name Display name of the AI Dataset Asset record.

Table: AI Dataset Digital Asset [alm_ai_dataset_digital_asset]

Data type: String
result.documentation Documentation of the associated AI Dataset Product Model [cmdb_ai_dataset_product_model] table record.

Data type: String
result.managed_by Details about the user who manages the AI Dataset Asset record.

Data type: Object

"managed_by": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.managed_by.name Name of the user who manages the AI Dataset Asset record.

Table: User [user]

Data type: String
result.managed_by.sys_id Sys_id of the user who manages the AI Dataset Asset record.

Table: User [user]

Data type: String
result.name Name of the associated AI Dataset Product Model record.

Table: AI Dataset Product Model [cmdb_ai_dataset_product_model]

Data type: String
result.provider Provider of the associated AI Dataset Product Model [cmdb_ai_dataset_product_model] table record.

Data type: Object

provider: {
  "name": "String",
  "sys_id": "String"
}
result.provider.name Name of the provider.

Data type: String
result.provider.sys_id Sys_id of the provider in the associated AI Dataset Product Model [cmdb_ai_dataset_product_model] table record.

Data type: String
result.source Details about the source of AI dataset asset.
Valid values:
  • Link to the source of the dataset asset.
  • Details (in plain text) of the source of the dataset asset. For example, the name of a product or website.

Data type: String

Default: empty or null
result.state State of the AI Dataset Asset record.
Possible values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String
result.sys_id Sys_id of the AI Dataset Asset record.

Table: AI Dataset Digital Asset [alm_ai_dataset_digital_asset]

Data type: String
result.updated Date and time that the AI Dataset Asset record was last updated.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.version Version number of the associated AI Dataset Product Model record. For example, V2.

Data type: String
result.warnings Comma-separated list of warning messages that are present when creating the dataset. These warnings can be validation checks, such as when the sys_id of an optional parameter is invalid.

Data type: Array

"warnings": ["String"]

Example: cURL request

The following example shows how to update details in an AI dataset record with a given ID. The request body contains the parameter values to update.

curl -X PUT 'https://instance.servicenow.com/api/sn_ent/asset/ai_dataset/9833721b331e92101c9aca989d5c7bf0' \ 
  -H 'Accept: application/json' \ 
  -H 'Content-Type: application/json' \ 
  -u 'username':'password' \ 
  -d '{ 
  "name": "Dataset One", 
  "description": "Description for dataset ", 
  "provider": "servicenow", 
  "version": "V1", 
  "state": 31, 
  “source”: “Source of dataset” 
  "documentation": "document", 
  “dataset_card”: “Dataset Card”, 
  “base_datasets”: [ “Dataset Two”, “Dataset Three”], 
  “data_type”: “1,2”, 
  “acceptable_usage”: “1,2”, 
  "managed_by": "abel.tuter" 
}'

Response body.

{ 
  "result": { 
    "asset": { 
      "sys_id": "da8393eb40d25210f877b00c113d1fc1", 
      "display_name": "ServiceNow Closed Incidents", 
      "name": "Closed Incidents", 
      "description": "Incidents with resolution", 
      "documentation": "Sample Documentation", 
      "source": "incident table on servicenow instance", 
      "dataset_card": "Dataset Card", 
      "state": "Deployed", 
      "version": null, 
      "data_type": { 
        "value": "1", 
        "label": "Text" 
      }, 
      "provider": { 
        "sys_id": "93d4ecfac0a8000b6294d71b733977fb", 
        "name": "ServiceNow" 
      }, 
      "managed_by": { 
        "sys_id": "undefined", 
        "name": "" 
      }, 
      "acceptable_usage": { 
        "value": "1", 
        "label": "Training" 
      }, 
      "base_datasets": [], 
      "created": "2024-12-12 01:23:03", 
      "updated": "2024-12-12 01:23:03" 
    }, 
    "warnings": [] 
       } 
}

AI Assets API - PUT /sn_ent/asset/ai_prompt/{sys_id}

Updates the data of a specific AI prompt asset record according to information you provide in the request body.

See also:
Note: Provide only the parameter-value pairs for the specific data you want to update. This endpoint overwrites the data for any parameters that are sent in the request.

URL format

Versioned URL: /api/sn_ent/{api_version}/asset/ai_prompt/{sys_id}

Default URL: /api/sn_ent/asset/ai_prompt/{sys_id}

Supported request parameters

Table 55. Path parameters
Name Description
api_version Optional. Version of the endpoint to access. For example, v1 or v2. Only specify this value to use an endpoint version other than the latest.

Data type: String
sys_id Sys_id of the asset in the AI Prompt Asset [alm_ai_prompt_digital_asset] table.

Data type: String
Table 56. Query parameters
Name Description
None
Table 57. Request body parameters (XML or JSON)
Name Description
object Required. Content of the AI Prompt asset to update.

Data type: Object

{
  "ai_model": "String",
  "description": "String",
  "documentation": "String",
  "managed_by": "String" 
  "name": "String",
  "prompt_info": "String", 
  "provider": "String",
  "state": Number,
  "version": "String"
}
ai_model Value of the AI Model field of an existing record in the AI Model Digital Asset [alm_ai_model_digital_asset] table.
Valid values:
  • Name of the AI Model Digital Asset [alm_ai_model_digital_asset] record
  • Sys_id of the AI Model Digital Asset [alm_ai_model_digital_asset] record

Data type: String

Default: empty string
description Description to give the AI Prompt Product Model.

Table: AI Prompt Product Model [cmdb_ai_prompt_product_model]

Data type: String

Default: empty string
documentation Documentation of the AI Prompt Product Model.

Data type: String

Default: empty string
managed_by Value of the Managed by field of an existing record in the User [sys_user] table.
Valid values:
  • Name of the User [sys_user] record
  • Sys_id of the User [sys_user] record

Data type: String

Default: empty string
name Required. Name of the associated record in the AI Prompt Product Model [cmdb_ai_prompt_product_model] table.

Data type: String

Default: empty string
prompt_info Prompt information for AI Prompt Asset.

Data type: String

Default: empty string
provider Required. Value of the Provider field of an existing record in the Company [core_company] table.
Valid values:
  • Name of the Company [core_company] record
  • Sys_id of the Company [core_company] record

Data type: String
state State to apply to the AI Model Asset. For example,
Valid values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String
version Version number to assign the AI Prompt Product Model. For example, V2.

Data type: String

Default: empty string

Headers

The following request and response headers apply to this HTTP action only, or apply to this action in a distinct way. For a list of general headers used in the REST API, see Supported REST API headers.

Table 58. Request headers
Header Description
Accept Data format of the response body. Only supports application/json.
Content-Type Data format of the request body. Only supports application/json.
Table 59. Response headers
Header Description
None

Status codes

The following status codes apply to this HTTP action. For a list of possible status codes used in the REST API, see REST API HTTP response codes.

Table 60. Status codes
Status code Description
200 Successful. The request was successfully processed.
401 Unauthorized. The user credentials are incorrect or have not been passed.
404 Not found. Failed to fetch the asset with the given sys_id.
500 Internal server error. An unexpected error occurred while processing the request. The response contains additional information about the error.

Response body parameters

Name Description
result 
result: {
  "asset": {Object},
  "warnings": [Array]
}
result.asset Details about the newly created entry.

Data type: Object

"asset": {
  "ai_model": Object,
  "created": "String",
  "description": "String",
  "display_name": "String",
  "documentation": "String",
  "managed_by": Object,
  "name": "String",
  "prompt_info": "String",
  "provider": Object,
  "state": "String",
  "sys_id": "String",
  "updated": "String",
  "version": "String",
  "warnings": [Array]
}
result.asset.ai_model List of AI models in the AI System Asset [cmdb_ai_ system_asset_model] table record.

Data type: Array

"ai_model": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.asset.ai_models.name Name of the AI System Digital Asset [alm_ai_system_digital_asset] table record.

Data type: String
result.asset.ai_models.sys_id Sys_id of the AI System Digital Asset [alm_ai_system_digital_asset] table record.

Data type: String
result.asset.created Date and time that the AI model asset was created.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.asset.description Description of the associated AI System Product Model record.

Table: AI System Product Model [cmdb_ai_system_product_model]

Data type: String
result.asset.display_name Display name of the AI System Asset record.

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: String
result.asset.documentation Documentation of the associated AI System Product Model [cmdb_ai_ system_product_model] table record.

Data type: String
result.asset.managed_by Details about the user who manages the asset.

Data type: Object

"managed_by": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.asset.managed_by.name Name of the user who manages the AI model asset record.

Table: User [user]

Data type: String
result.asset.managed_by.sys_id Sys_id of the user who manages the AI model asset record.

Table: User [user]

Data type: String
result.asset.name Name of the associated AI System Product Model record.

Table: AI System Product Model [cmdb_ai_system_product_model]

Data type: String
result.asset.prompt_info Prompt Information of the AI Prompt Asset record.

Data type: String
result.asset.provider Value of the Provider field in the associated AI System Product Model [cmdb_ai_ system_product_model] table record.

Data type: Object

provider: {
  "name": "String",
  "sys_id": "String"
}
result.asset.provider.name Name of the provider in the associated AI System Product Model [cmdb_ai_ system_product_model] table record.

Data type: String
result.asset.provider.sys_id Sys_id of the provider in the associated AI System Product Model [cmdb_ai_ system_product_model] table record.

Data type: String
result.asset.state State of the AI Model Asset.
Possible values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String
result.asset.sys_id Sys_id of the AI System Asset record.

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: String
result.asset.updated Date and time that the AI model asset was last updated.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.asset.version Version number of the associated AI System Product Model record. For example, V2.

Data type: String
result.warning Comma-separated list of warning messages. These warnings can be validation checks, such as when the sys_id of an optional parameter is invalid.

Data type: Array

"warnings": ["String"]

Example: cURL request

The following example updates the data of the AI Prompt asset with the given sys_id according to the provided parameter values in the request body.

curl -X PUT 'https://instance.servicenow.com/api/sn_ent/asset/ai_prompt/9833721b331e92101c9aca989d5c7bf0' \
  -H 'Accept: application/json' \
  -H 'Content-Type: application/json' \
  -u 'username':'password' \
  -d '{
  "name": "Incident Summarization prompt1",
  "description": "Prompt for Incident Summarization",
  "provider": "servicenow",
  "version": "V1",
  "state": 31,
  "documentation": "Docuuu",
  "ai_model": "mixtral-instruct",
  "prompt_info": "Provide incident summary using short_decription, state, worknotes",
  "managed_by": "abel.tuter"
 }'

Response body:

{
  "result": {
    "asset": {
      "sys_id": "9833721b331e92101c9aca989d5c7bf0",
      "display_name": "ServiceNow Incident Summarization prompt1 V1",
      "name": "Incident Summarization prompt1",
      "description": "Prompt for Incident Summarization",
      "version": "V1",
      "provider": {
        "sys_id": "93d4ecfac0a8000b6294d71b733977fb",
        "name": "ServiceNow"
      },
      "documentation": "Docuuu",
      "state": "Deployed",
      "ai_model": {
        "sys_id": "9d7dc7e6eb1e5210aa82fab8bad0cda2",
        "name": "mixtral-instruct"
      },
      "prompt_info": "Provide incident summary using short_decription, state, worknotes",
      "managed_by": {
        "sys_id": "62826bf03710200044e0bfc8bcbe5df1",
        "name": "Abel Tuter"
      },
      "created": "2024-12-11 04:23:17",
      "updated": "2024-12-11 04:23:17"
    },
    "warnings": []
  }
}

AI Assets API - PUT /sn_ent/asset/ai_system/{sys_id}

Updates the data of a specific AI system record according to information you provide in the request body.

Note: Provide only the parameter-value pairs for the specific data you want to update. This endpoint overwrites the data for any parameters that are sent in the request.

URL format

Versioned URL: /api/sn_ent/{api_version}/asset/ai_system/{sys_id}

Default URL: /api/sn_ent/asset/ai_system/{sys_id}

Supported request parameters

Table 61. Path parameters
Name Description
api_version Optional. Version of the endpoint to access. For example, v1 or v2. Only specify this value to use an endpoint version other than the latest.

Data type: String
sys_id Sys_id of the asset in the AI System Digital Asset [alm_ai_system_digital_asset] table.

Data type: String
Table 62. Query parameters
Name Description
None
Table 63. Request body parameters (XML or JSON)
Name Description
object Required. Details to update in the AI system.

Data type: Object

{ 
  "name": "String", 
  "description": "String", 
  "provider": "String", 
  "version": "String", 
  "state": Number, 
  "documentation": "String", 
  "ai_models": "String", 
  "ai_prompts": "Strings", 
  "evaluation_datasets": [Array], 
  "evaluation_metrics_report": "String", 
  "managed_by": "String" 
}
object.name Required. Name for the AI System Product Model.

Data type: String
object.description Description for the AI System Product Model.

Data type: String
object.provider Required. Value of the Provider field in an existing Company [core_company] table record.
Valid values:
  • Name of the Company [core_company] record
  • Sys_id of the Company [core_company] record

Data type: String
object.state State to apply to the AI System Asset.
Valid values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String

Default: empty string
object.version Version for AI System Product Model. For example, V2.

Data type: String

Default: empty string
object.documentation Documentation of the AI System Asset.

Data type: String

Default:
object.ai_models Comma-separated list of strings, where each string represents an AI Model field value of an existing AI Model Digital Asset record.

Table: AI Model Digital Asset [alm_ai_model_digital_asset]

Valid values:
  • Name of the AI Model Digital Asset record.
  • Sys_id of the AI Model Digital Asset record.

Data type: String

Default: empty string
object.evaluation_datasets Comma-separated list of AI datasets used for evaluating the AI System Asset record. Mostly applicable for models developed within an organization.

Tables: AI Dataset Digital Asset [alm_ai_dataset_digital_asset], AI System Asset [AI System Digital Asset alm_ai_system_digital_asset]

Data type: Array

"evaluation_datasets": [{String", "String"}]
Valid values:
  • Name of the AI Dataset Digital Asset record.
  • Sys_id of the AI Dataset Digital Asset record.

Default: empty string
object.evaluation_metrics_report Reference to the evaluation results located within the text field of the AI system asset record. For example:
Testing results:  link to the result document

Details:
Accuracy: 85%
Hallucination: 10%
Eval Run 1: link
Eval Run 2: link

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: String

Default: empty string
object.managed_by Value of the 'Managed by' field of an existing User [sys_user] table record.
Valid values:
  • Name of the User [sys_user] record
  • Sys_id of the User [sys_user] record

Data type: String

Default: empty string

Headers

The following request and response headers apply to this HTTP action only, or apply to this action in a distinct way. For a list of general headers used in the REST API, see Supported REST API headers.

Table 64. Request headers
Header Description
Accept Data format of the response body. Only supports application/json.
Content-Type Data format of the request body. Only supports application/json.
Table 65. Response headers
Header Description
None

Status codes

The following status codes apply to this HTTP action. For a list of possible status codes used in the REST API, see REST API HTTP response codes.

Table 66. Status codes
Status code Description
200 Successful. The request was successfully processed.
400 Bad Request. A bad request type or malformed request was detected.
404 Not found. Failed to fetch the asset with the given sys_id.
500 Internal server error. An unexpected error occurred while processing the request. The response contains additional information about the error.

Response body parameters (JSON or XML)

Name Description
result Results of the updated AI System Asset.

Data type: Object

"result": {
  "asset": Object,
  "warnings": [Array]
}
result.asset Details about the newly created asset.

Data type: Object

"asset": {
  "ai_model": [Array],
  "ai_prompts": [Array]
  "created": "String",
  "description": "String",
  "display_name": "String",
  "documentation": "String",
  "evaluation_datasets": Array,
  "evaluation_metrics_report": String,
  "managed_by": Object,
  "name": "String",
  "provider": Object,
  "state": "String",
  "sys_id": "String",
  "updated": "String",
  "version": "String",
  "warnings": [Array]
}
result.ai_models List of AI models in the AI System Digital Asset [alm_ai_system_digital_asset] table record.

Data type: Array

"ai_models": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.ai_models.name Name of the AI System Digital Asset record.

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: String
result.ai_models.sys_id Sys_id of the AI System Digital Asset record.

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: String
result.ai_prompts List of AI Prompts in the AI System Asset record.

Data type: Array

"ai_prompts": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.ai_prompts.name Name of the AI prompt.

Data type: String
result.ai_prompts.sys_id Sys_id of the AI Prompt Digital Asset record.

Table: AI Prompt Digital Asset [alm_ai_prompt_digital_asset]

Data type: String
result.asset.created Date and time that the AI model asset was created.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.asset.description Description of the associated AI System Product Model record.

Table: AI System Product Model [cmdb_ai_system_product_model]

Data type: String
result.asset.display_name Display name of the AI System Asset record.

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: String
result.asset.documentation Documentation of the associated AI System Product Model [cmdb_ai_ system_product_model] table record.

Data type: String
result.asset.evaluation_datasets List of AI datasets used for evaluating the AI System Asset record. Mostly applicable for models developed within an organization.

Tables: AI Dataset Digital Asset [alm_ai_dataset_digital_asset], AI System Asset [alm_ai_system_digital_asset]

Data type: Array

"evaluation_datasets": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.asset.evaluation_datasets.name Name of the AI Dataset Digital Asset.

Data type: String
result.asset.evaluation_datasets.sys_id Sys_id of the AI Dataset Digital Asset [alm_ai_dataset_digital_asset] table record.

Data type: String
result.asset.evaluation_metrics_report Evaluation results of the AI system asset.
Possible values:
  • Details (in plain text) outlining results
  • Links to specific results

Data type: String
result.asset.managed_by Comma-separated list of sys_ids or display names of AI datasets of the AI System Asset used for evaluating the model. Mostly applicable for models developed within an organization.

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: Object

"managed_by": [
  {
  "name": "String",
  "sys_id": "String"
  }
]
result.asset.managed_by.name Sys_id of the user who manages the AI model asset record.

Table: User [user]

Data type: String
result.asset.managed_by.sys_id Name of the user who manages the AI model asset record.

Table: User [user]

Data type: String
result.asset.name Name of the associated AI System Product Model record.

Table: AI System Product Model [cmdb_ai_system_product_model]

Data type: String
result.asset.provider Value of the Provider field in the associated AI System Product Model [cmdb_ai_ system_product_model] table record.

Data type: Object

provider: {
  "name": "String",
  "sys_id": "String"
}
result.asset.provider.name Name of the provider in the associated AI System Product Model [cmdb_ai_ system_product_model] table record.

Data type: String
result.asset.provider.sys_id Sys_id of the provider in the associated AI System Product Model [cmdb_ai_ system_product_model] table record.

Data type: String
result.asset.state State of the AI Model Asset.
Possible values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String
result.asset.sys_id Sys_id of the AI System Asset record.

Table: AI System Digital Asset [alm_ai_system_digital_asset]

Data type: String
result.asset.updated Date and time that the AI model asset was last updated.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.asset.version Version number of the associated AI System Product Model record. For example, V2.

Data type: String
result.asset.warnings Comma-separated list of warning messages. These warnings can be validation checks, such as when the sys_id of an optional parameter is invalid.

Data type: Array

"warnings": ["String"]

Example: cURL request

The following example updates the AI System with details to update in the request body.

curl -X PUT 'https://instance.servicenow.com/api/sn_ent/asset/ai_prompt/3b140397435a9210a63d00002fb8f2d7' \ 
  -H 'Accept: application/json' \ 
  -H 'Content-Type: application/json' \ 
  -u 'username':'password' \ 
  -d '{ 
  "name": "Incident Summarization", 
  "description": "Skill to summarize incident", 
  "provider": "servicenow", 
  "documentation": "Sample Documentation", 
  "version": "V3", 
  "state": 31, 
  "ai_models": [ 
    "llm_generic_small" 
  ], 
  "ai_prompts": [ 
    "LLM Prompt" 
  ], 
  "evaluation_datasets": [ 
    "Base dataset" 
  ], 
  "evaluation_metrics_report": "Sample Report", 
  "managed_by": "abel.tutor" 
}'

Response body shows the results of the update that was applied to the given AI system sys_id.

{ 
  "result": { 
    "asset": { 
      "sys_id": "3b140397435a9210a63d00002fb8f2d7", 
      "display_name": "ServiceNow Incident Summarization V2", 
      "name": "Incident Summarization", 
      "description": "Skill to summarize incident", 
      "version": "V3", 
      "provider": { 
        "sys_id": "93d4ecfac0a8000b6294d71b733977fb", 
        "name": "ServiceNow" 
      }, 
      "documentation": "Sample Documentation", 
      "state": "Deployed", 
      "ai_models": [{  
       "sys_id": "9tgdc7e6eb1e5210aa82fab8bad0cda2",  
       "name": "llm_generic_small"  
     }], 
      "ai_prompts": [{  
       "sys_id": "7d7dc7e6eb1e5210aa82fab8bad0cda2",  
       "name": "LLM Prompt"  
     }], 
      "evaluation_datasets": [{  
       "sys_id": "9d7dc7e6eb1e5210aa82fab8bad0cda2",  
       "name": "Base dataset"  
     }], 
      "evaluation_metrics_report": "Sample Report", 
      "managed_by": { 
        "sys_id": "a8f98bb0eb32010045e1a5115206fe3a", 
        "name": "Abraham Lincoln" 
      }, 
      "created": "2024-12-11 19:07:13", 
      "updated": "2024-12-11 19:07:42" 
    }, 
    "warnings": [] 
  } 
}

AI Assets API - PUT /sn_ent/asset/ai_model/{sys_id}

Updates the data of a specific AI model asset record according to information you provide in the request body.

Note: Provide only the parameter-value pairs for specific data you want to update. This endpoint overwrites the data for all parameters that are sent in the request.

URL format

Versioned URL: /api/sn_ent/{api_version}/asset/ai_model/{sys_id}

Default URL: /api/sn_ent/asset/ai_model/{sys_id}

Supported request parameters

Table 67. Path parameters
Name Description
api_version Optional. Version of the endpoint to access. For example, v1 or v2. Only specify this value to use an endpoint version other than the latest.

Data type: String
sys_id Sys_id of the asset in the AI Prompt Asset [alm_ai_prompt_digital_asset] table.

Data type: String
Table 68. Query parameters
Name Description
None
Table 69. Request body parameters (XML or JSON)
Name Description
object Required. Details to update in the given AI model.

Data type: Object

{
  "base_model": "String",
  "context_window": "String",
  "deployment_guideline": "String",
  "description": "String",
  "documentation": "String",
  "evaluation_datasets": "String",
  "evaluation_metrics_report": "String",
  "managed_by": "String",
  "model_size_in_mb": "String",
  "model_weights_info": "String",
  "name": "String",
  "parameters_info": "String",
  "provider": "String",
  "required_infrastructure": "String",
  "state": Number,
  "supported_languages": "String",
  "training_datasets": "String",
  "training_procedure": "String",
  "version": "String"
}
base_model AI model that this model version was derived from.
Note: Only applicable for models developed within the organization.

Data type: Object

{
  "name": "String",
  "sys_id": "String"
 }

Default: empty object
base_model.name Name of the AI model asset to model this AI model after.

Data type: String
base_model.sys_id Sys_id of the AI model asset to model this AI model after.

Table: AI Model Digital Asset alm_ai_model_digital_asset

Data type: String
context_window Size of input sequences (in other words, the number of tokens) that the model can handle.

Data type: Integer

Default: 0
deployment_guideline Instructions applicable for models developed and deployed within an organization.

Data type: String

Default: empty
description Description to give the AI Model Product Model.

Updated in table: AI Model Product Model [cmdb_ai_model_product_model]

Data type: String

Default: empty
documentation Documentation of the AI Prompt Product Model record.

Table: AI Prompt Product Model [cmdb_ai_model_product_model]

Data type: String

Default: empty
evaluation_datasets Comma-separated list of sys_ids or display names of AI datasets of the AI Prompt Digital Asset used for evaluating the model. Mostly applicable for models developed within an organization.

Tables: AI Dataset Digital Asset [alm_ai_dataset_digital_asset], AI Prompt Digital Asset [alm_ai_prompt_digital_asset]

Data type: Array

"evaluation_datasets": [
  "String",
  "String"
]
Valid values:
  • Name of the AI Dataset Digital Asset record.
  • Sys_id of the AI Dataset Digital Asset record.

Default: empty string
evaluation_metrics_report Reference to the evaluation results located within the text field of the AI Dataset Asset record. For example:
Testing results:  link to the result document

Details:
Accuracy: 85%
Hallucination: 10%
Eval Run 1: link
Eval Run 2: link

Table: AI Dataset Asset [alm_ai_dataset_digital_asset]

Data type: String

Default: empty string
managed_by Value of the 'Managed by' field of an existing record in the User [sys_user] table.
Valid values:
  • Name of the User [sys_user] record
  • Sys_id of the User [sys_user] record

Data type: String

Default: empty
model_size_in_mb Size of the model in MB. Mostly applicable for models developed and deployed within an organization.

Data type: Number

Default: empty
model_weights_info Additional model information, if available. Mostly applicable for models developed within an organization.

Data type: String

Default: empty
name Required. Name of the AI Model Product Model.

Updated in table: AI Model Product Model [cmdb_ai_model_product_model]

Data type: String
parameters_info Number of parameters to give to the model.

Data type: String

Default: empty string
provider Required. Value of the 'Provider' field of an existing Company [core_company] table record. Assigns this provider to the new AI model.
Valid values:
  • Name of the Company [core_company] record
  • Sys_id of the Company [core_company] record

Data type: String
required_infrastructure Documentation of infrastructure needs for the model deployment. For example, details about the infrastructure stack and processing needs. Mostly applicable for models deployed within an organization.

Data type: String

Default: empty
source Details about the source of who or what created the model.
Valid values:
  • Link to the source of the model. For example, https://huggingface.co/mistralai/model1 or a link to Github.
  • Details of the source in plain text. For example, Microsoft Azure

Data type: String

Default: empty or null
state State to apply to the AI Model Asset.
Valid values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String
supported_languages Comma-separated list of languages that are supported by the AI Model.

Data type: Array

"supported_languages": [
  "String",
  "String"
]
Valid values:
  • Sys_id of the Language [sys_language] record
  • Name of the Language [sys_language] record. For example, "French", "English"

Default: empty string
training_datasets Reference to 1+ associated datasets used for training the model. Comma-separated list of sys_ids or display names of the AI Dataset Asset [alm_ai_dataset_digital_asset] table. Mostly applicable for models developed within an organization.

Data type: Array

"training_datasets": [
  "String",
  "String"
]

Default: empty string
training_procedure Type of training to apply to the AI model.
Valid values:
  • 1: Decision Trees
  • 2: Deep Neural Networks
  • 3: Linear Regression
  • 4: Logistic Regression
  • 5: Random Forest
  • 6: Supervised Learning
  • 7: Unsupervised Learning
  • 8: Reinforcement Learning
  • 9: Transfer Learning
  • 10: Semi-Supervised Learning
  • 11: Instruction Finetuning
  • 12: Supervised Finetuning

Data type: String represented with a number (for example, "3")

Default: 1: Decision Trees
version Version for AI Model Product Model. For example, V2.

Data type: String

Default: empty

Headers

The following request and response headers apply to this HTTP action only, or apply to this action in a distinct way. For a list of general headers used in the REST API, see Supported REST API headers.

Table 70. Request headers
Header Description
Accept Data format of the response body. Supported types: application/json or application/xml.

Default: application/json
Table 71. Response headers
Header Description
None

Status codes

The following status codes apply to this HTTP action. For a list of possible status codes used in the REST API, see REST API HTTP response codes.

Table 72. Status codes
Status code Description
200 Successful. The request was successfully processed.

Response body parameters (JSON or XML)

Name Description
result Results of the AI model asset creation.

Data type: Object

"result": { 
    "asset": Object, 
    "warnings": [Array], 
 }
result.asset Details about the created AI model asset.

Data type: Object

"asset": {
  "base_model": Object,
  "context_window": String,
  "created": "String",
  "deployment_guideline": String,
  "description": "String",
  "display_name": "String",
  "documentation": "String",
  "evaluation_datasets": Array,
  "evaluation_metrics_report": String,
  "managed_by": Object,
  "model_size_in_mb": String,
  "name": "String",
  "parameters_info": "String",
  "provider": Object,
  "required_infrastructure": String,
  "state": "String",
  "source": String, 
  "supported_languages": Array,
  "sys_id": "String",
  "training_datasets": Array,
  "training_procedure": String,
  "updated": "String",
  "version": "String"
}
result.asset.base_model Information about the AI model asset from which this version derived from.

Data type: Object

"base_model": { 
  "name": "String",
  "sys_id": "String"
}
result.asset.base_model.name Display name of AI model asset.

Table: AI Model Digital Asset [alm_ai_model_digital_asset]

Data type: String
result.asset.base_model.sys_id Sys_id of AI model asset.

Table: AI Model Digital Asset [alm_ai_model_digital_asset]

Data type: String
result.asset.context_window Size of input sequences that the model can handle. In other words, the number of tokens.

Data type: String represented with a number. For example, "6000".
result.asset.created Date and time that the AI model asset was created.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.asset.deployment_guideline Instructions applicable for models developed and deployed within an organization.

Data type: String
result.asset.description Description of the associated AI Model Product Model record.

Table: AI Model Product Model [cmdb_ai_model_product_model]

Data type: String
result.asset.display_name Display name of the AI Model Asset record.

Table: AI Model Digital Asset [alm_ai_model_digital_asset] (display_name field)

Data type: String
result.asset.documentation Documentation of the associated AI Model Product Model record.

Table: AI Model Product Model [cmdb_ai_model_product_model]

Data type: String
result.asset.evaluation_datasets Comma-separated list of sys_ids or display names of AI datasets of the AI Model Digital Asset used for evaluating the model. Mostly applicable for models developed within an organization.

Tables: AI Dataset Digital Asset [alm_ai_dataset_digital_asset], AI Model Digital Asset [alm_ai_model_digital_asset]

Data type: Array

"evaluation_datasets": [
  { 
  "name": "String",
  "sys_id": "String"
  } 
]
result.evaluation_datasets.name Name of the AI Dataset Digital Asset.

Data type: String
result.evaluation_datasets.sys_id Sys_id of the AI Model Digital Asset record.

Table: AI Model Digital Asset [alm_ai_model_digital_asset]

Data type: String
result.asset.evaluation_metrics_report Reference to the evaluation results.
Possible values:
  • Details (in plain text) outlining results
  • Links to specific results

Data type: String
result.asset.managed_by User that manages the AI model asset record.

Data type: Object

" managed_by": {
  "name": "String",
  "sys_id": "String"
}
result.asset.managed_by.name Name of the user who manages the AI model asset record.

Table: User [user]

Data type: String
result.asset.managed_by.sys_id Sys_id of the user who manages the AI model asset record.

Table: User [user]

Data type: String
result.asset.model_size_in_mb Size of the model in MB. Usually applicable for models developed and deployed within an organization.

Data type: Number
result.asset.name Name of the associated AI Model Product Model record.

Table: AI Model Product Model [cmdb_ai_model_product_model]

Data type: String
result.asset.provider Provider of the associated AI Model Product Model record.

Table: AI Model Product Model [cmdb_ai_model_product_model]

Data type: Object

"provider": {
  "name": "String",
  "sys_id": "String"
}
result.asset.provider.name Name of the provider.

Data type: String
result.asset.provider.sys_id Sys_id of the record from Company [core_company] table that corresponds to the Provider of the associated AI Model Product Model record.

Table: AI Model Product Model [cmdb_ai_model_product_model]

Data type: String
result.asset.source Details about the source of the asset.
Valid values:
  • Link to the source of the model. For example, https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1 or a link to Github.
  • Details of the source in plain text. For example, Microsoft Azure

Data type: String
result.asset.state State of the AI Model Asset record.
Possible values:
  • 1: In use
  • 31: Deployed
  • 32: Retired
  • 33: Development
  • 34: Unknown
  • 35: N/A

Data type: String
result.asset.supported_languages.name Name of the supported language.

Table: Language [sys_language]

Data type: String
result.asset.supported_languages.sys_id Sys_id of the supported language.

Table: Language [sys_language]

Data type: String
result.asset.sys_id Sys_id of the AI Model Asset record

Table: AI Model Digital Asset [alm_ai_model_digital_asset]

Data type: String
result.asset.training_datasets Reference to 1+ associated data sets used for training the model.

Data type: Array

"training_datasets": [
  { 
  "name": "String",
  "sys_id": "String"
  } 
]
result.asset.training_procedure Type of AI training applied to the model.
Possible values:
  • 1: Decision Trees
  • 2: Deep Neural Networks
  • 3: Linear Regression
  • 4: Logistic Regression
  • 5: Random Forest
  • 6: Supervised Learning
  • 7: Unsupervised Learning
  • 8: Reinforcement Learning
  • 9: Transfer Learning
  • 10: Semi-Supervised Learning
  • 11: Instruction Finetuning
  • 12: Supervised Finetuning

Data type: String
result.asset.updated Date and time that the AI model asset was last updated.

Format: YYYY-MM-DD HH:mm:ss

Data type: String
result.asset.version Version of the associated AI Model Product Model record.

Table: AI Model Product Model [cmdb_ai_model_product_model table]

Data type: String
result.required_infrastructure Documentation of infrastructure needs for the model's deployment. For example, details about the infrastructure stack and processing needs.

Data type: String
result.asset.parameters_info Properties of the training data that learn during the learning process. For example: 7B or 30B.

Data type: String
result.asset.supported_languages Details about the languages supported by the AI model asset.

Table: Language [sys_language]

Data type: Object

"supported_languages": [
 { 
  "name": "String",
  "sys_id": "String"
  }
]
result.warnings Comma-separated list of warning messages. These warnings can be validation checks, such as when the sysId of an optional parameter is invalid.

Data type: Array

"warnings": ["String"]

Example: cURL request

The following example shows how to update a given AI asset model using the PUT method.

curl -X PUT 'https://instance.servicenow.com/api/sn_ent/asset/ai_model/9833721b331e92101c9aca989d5c7bf0' \ 
  -H 'Accept: application/json' \ 
  -H 'Content-Type: application/json' \ 
  -u 'username':'password' \ 
-d '{ 
    "name": "Now LLM", 
    "description": "enables text-to-text like question answering and summarization", 
    "provider": "servicenow", 
    "documentation": "Now LLM V5 Documentation", 
    "version": "V8", 
    "parameters_info": "7B", 
    "supported_languages": [ 
        "English", 
        "French" 
    ], 
    "model_size_in_mb": "87", 
    "deployment_guideline": "Deployed on ServiceNow infrastructure", 
    "source": "huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1", 
    "training_procedure": "2", 
    "context_window": "8000", 
    "state": "31", 
    "base_model": "Servicenow Now LLM V6", 
    "model_weights_info": "refer to weights and biases project", 
    "required_infrastructre": "GPUs needed: 1, GPU Type: A100", 
    "training_datasets": [ 
        "Servicenow Open Incidents" 
    ], 
    "evaluation_datasets": [ 
        "Servicenow Open Incidents" 
    ], 
    "evaluation_metrics_report": "Testing results: See files attached to this record", 
    "managed_by": "abel.tuter" 
}'

Response body.

 { 
  "result": { 
    "asset": { 
      "sys_id": "a438d170ff96da10c1fbffffffffffd5", 
      "display_name": "ServiceNow Now LLM V6", 
      "name": "Now LLM", 
      "description": "enables text-to-text like question answering and summarization", 
      "version": "V8", 
      "provider": { 
        "sys_id": "93d4ecfac0a8000b6294d71b733977fb", 
        "name": "ServiceNow" 
      }, 
      "documentation": "Now LLM V5 Documentation", 
      "parameters_info": "7B", 
      "supported_languages": [ 
        { 
          "sys_id": "914493a30f320010e96b0e4fef767e90", 
          "name": "English" 
        } 
      ], 
      "model_size_in_mb": "87", 
      "deployment_guideline": "Deployed on ServiceNow infrastructure", 
      "source": "https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1", 
      "training_procedure": "2", 
      "context_window": "8000", 
      "state": "Deployed", 
      "required_infrastructure": "undefined", 
      "base_model": { 
        "sys_id": "a438d170ff96da10c1fbffffffffffd5", 
        "name": "ServiceNow Now LLM V8" 
      }, 
      "evaluation_datasets": [ 
        { 
          "sys_id": "45cb45baff06d610c1fbffffffffffa9", 
          "name": "ServiceNow Open Incidents" 
        } 
      ], 
      "training_datasets": [ 
        { 
          "sys_id": "45cb45baff06d610c1fbffffffffffa9", 
          "name": "ServiceNow Open Incidents" 
        } 
      ], 
      "evaluation_metrics_report": "Testing results: See files attached to this record", 
      "managed_by": { 
        "sys_id": "62826bf03710200044e0bfc8bcbe5df1", 
        "name": "Abel Tuter" 
      }, 
      "created": "2024-12-03 16:50:53", 
      "updated": "2024-12-12 15:56:28" 
    }, 
    "warnings": [ 
      "Reference record 'French' not found for supported_languages in table sys_language" 
    ] 
  } 
}