Is there any out-of-the-box AI or GenAI features available within ServiceNow?

[srinivas_366]

Is there an out-of-the-box ServiceNow AI/GenAI capability that can analyze inbound email subject and body content and automatically determine whether to create an Incident or a Generic Service Request?

Any experience with Now Assist, AI Agents, or other OOTB solutions would be appreciated.

[Dr Atul - LNG]

No, there is an OOTB skill available for this, so I don’t think we need to build it from scratch on our side.

You can use a GenAI Action/Flow to read the data and then make a decision based on the provided record. @srinivas_366 

[srinivas_366]

@Dr Atul - LNG 

If possible could you please give me brief explanation?

[Dr Atul - LNG]

Hi @srinivas_366 

I think my earlier post was a little too quick — I’ve since found something even more interesting! 😄

https://www.servicenow.com/docs/r/yokohama/platform-administration/email-agentic-workflow.html?conte...

something i got from AI: Please check the links as well

Yes, ServiceNow provides out-of-the-box AI capabilities designed to analyze inbound emails and automatically classify them. However, the specific tool you use depends on whether you utilize traditional Machine Learning (ML) or the newest Generative AI / Agentic architecture. [1, 2]

ServiceNow provides these specific capabilities across two distinct frameworks:

1. The GenAI & Agentic Approach (Email Agentic Workflow)

Introduced in recent releases (such as Yokohama/Brazil) as part of the ServiceNow AI Platform, the platform includes a native Email Agentic Workflow. [1, 2]

• How it works: This feature uses specialized AI Agents (driven by LLMs). The Intent Identification Agent reads the natural language context of the email subject and body to determine the user's ultimate goal. [1]

• The Routing: The Intent Executor Agent maps that identified intent to an action—such as triggering an automated subflow to create an Incident or initiating a specific Service Request. This removes the need for brittle, regex-based keyword rules. [1, 2]

2. The Traditional ML Approach (Predictive Intelligence)

If you are looking for classification that doesn't rely strictly on Generative LLMs, ServiceNow’s classic out-of-the-box feature is Predictive Intelligence. [1, 2]

• How it works: It uses an internal Classification Framework trained on your historical ticket data.

• The Routing: You can route all incoming emails to a staging table (or default them to an Incident). Predictive Intelligence evaluates the short description and body text to determine if the ticket belongs in the Incident [incident] table or should be reclassified into a Service Request [sc_req_item]. [1, 2, 3, 4, 5]

Comparison of the Two OOTB Approaches

Feature Capability	Email Agentic Workflow (GenAI)	Predictive Intelligence (ML)
Logic Type	Contextual reasoning & Intent matching via Large Language Models (LLMs).	Pattern matching & Statistical probability based on your text history.
Training Required	Zero training required; works out-of-the-box using natural language understanding.	Requires a historical data training set of records to learn from.
Setup Tool	Workflow Studio / Now Assist Admin.	Predictive Intelligence Application Workbench.

If you are licensed for Now Assist / ServiceNow Otto, the Email Agentic Workflow is the recommended path for intent-based classification. If you are on standard ITSM licenses, Predictive Intelligence is the standard mechanism to achieve this. [1, 2, 3]