---
sourceDocument: Brazil Enable AI
sourceDocumentLink: https://www.servicenow.com/docs/r/intelligent-experiences

 Release :

    - brazil

ft:locale :

    - en-US

ft:publication_title :

    - Brazil Enable AI

ft:clusterId :

    - platai

bundleId :

    - platai

workflow :

    - Platform


---

# Bring your own key for third-party AI provider integration

# Bring your own key for third-party AI provider integration {#ariaid-title1}

Release version: Brazil  
Updated September 10, 2026  
![](https://www.servicenow.com/docs/portal-asset/ico-clock) 2 minutes to read
Summarize  
![AI sparkle icon](https://servicenow.com/docs/portal-asset/ai-sparkle-icon) Summarized using AI  
This content was generated using new OpenAI-powered functionality. Results are provided on an as is basis and are not guaranteed to be accurate or complete.  

## Summary of Bring your own key for third-party AI provider integration

The Bring Your Own Key (BYOK) feature allows ServiceNow customers to use their own API credentials from supported cloud AI providers---such as Azure OpenAI, Amazon Bedrock (Anthropic Claude), and Google Gemini---to run generative AI skills and AI agents.
When BYOK is enabled, AI model requests are processed through the customer's cloud AI provider account instead of the default ServiceNow AI Platform infrastructure.
This empowers customers to retain control of their AI configuration, security, and data processing environment while still leveraging ServiceNow's AI capabilities.
Show full answer Show less  
BYOK supports the same range of AI models available on the ServiceNow AI Platform, ensuring compatibility and seamless integration. However, customers should be aware that using their own cloud AI provider account might incur additional charges from those providers.

## Key Features

* Integration with major cloud AI providers: Azure OpenAI, Amazon Bedrock (Anthropic Claude), and Google Gemini.
* API credential management remains under customer control, enhancing security and compliance.
* Supports mapping of ServiceNow AI model aliases to specific provider model versions to maintain consistency in AI skill behavior.
* Ability to override AI skill variants to customize AI behavior per deployment.
* Validation via outbound HTTPS logs to ensure requests are processed using the customer's own AI provider account.
* Detailed configuration workflows tailored per provider, including creation of connections, credentials, and model deployment settings.

## Practical Implementation for ServiceNow Customers

To enable BYOK, customers must configure a connection between their ServiceNow instance and their chosen cloud AI provider. This involves:

* Creating connection credentials such as API keys, endpoints, and region identifiers.
* Enabling BYOK in the ServiceNow AI Admin Hub for the selected provider.
* Mapping ServiceNow model aliases to specific AI model versions deployed in the customer's cloud account.
* Overriding model variants within AI skills through Version Management.
* Verifying that AI requests utilize the customer's own credentials via outbound HTTPS logging.

For Azure OpenAI specifically, customers need to configure API credentials, specify model and deployment names within the Generative AI Controller, and can customize prompt configurations for their large language models (LLMs). Switching to BYOK for AI processing is managed through integration settings.

## Why This Matters

BYOK enables ServiceNow customers to maintain full ownership and control over AI data processing and security by using their own cloud AI provider accounts. This is crucial for organizations with strict compliance, security, or data residency requirements. By integrating their own keys, customers can leverage ServiceNow's AI platform capabilities while aligning with their internal policies and managing potential costs directly with their chosen cloud AI provider.  
The bring-your-own-key (BYOK) feature enables you to use your own API credentials from supported cloud AI providers, such as Azure OpenAI, Amazon Bedrock, and Google Gemini, to run generative AI skills and AI agents.

When BYOK is enabled, AI model requests are sent through your cloud AI provider account instead of ServiceNow AI Platform infrastructure. This approach lets you maintain ownership of your AI configuration, security controls, and data processing environment while continuing to use ServiceNow AI Platform capabilities.  
BYOK supports the same AI models that the ServiceNow AI Platform supports. After you configure your API credentials, your provider resources are used to process requests instead of the default ServiceNow AI provider. BYOK keeps your keys under your control and AI workloads running within your cloud environment.  
Note:  
Using your organization's cloud AI provider account might result in additional charges from that provider. Review the provider's pricing and terms of service before enabling BYOK.

## ServiceNow integrated model providers {#byok-for-azure-open-ai__section_njw_gt2_zhc}

BYOK supports the following cloud AI providers:

* Azure OpenAI
* Anthropic Claude, which is accessible through Amazon Bedrock
* Google Gemini
{#byok-for-azure-open-ai__ul_pxb_kt2_zhc}

## Configuring a BYOK connection with a cloud AI provider {#byok-for-azure-open-ai__section_yqb_py3_yhc}

To use your cloud AI provider resources within your ServiceNow instance, configure a connection between ServiceNow and your provider. Configuration steps vary by provider, but follow a similar high-level configuration workflow, such as the following:

1. Create connection and credentials for the provider, such as endpoint, API key, and region.
2. Enable BYOK for the provider in AI Admin Hub.
3. Map ServiceNow model aliases to provider model versions.

   For example, map `gemini_large` → 2.5 Pro.
4. Override variants in each skill in Version Management.
5. Validate that requests are using your AI provider account by reviewing outbound HTTPS logs.
{#byok-for-azure-open-ai__ol_fdb_kwh_b3c}

For step-by-step instructions for each supported cloud AI provider, see [Configure Bring Your Own Key (BYOK) for Now Assist and AI Agents \[KB2666298\]](https://support.servicenow.com/kb?id=kb_article_view&sysparm_article=KB2666298) article in the Now Support
Knowledge Base.  

## High-level BYOK configuration for Azure OpenAI {#byok-for-azure-open-ai__example_grf_rb3_b3c}

1. Set up Azure OpenAI connection with your own key to use Azure OpenAI within your ServiceNow instance.

   For more information, see [Configure API credentials for Azure OpenAI](https://www.servicenow.com/docs/FEhu~qjmXYSbggefW_wZpw "Configure your API credentials to use Azure OpenAI as your LLM provider for Generative AI Controller capabilities.").
2. Configure your Azure OpenAI model and deployment name in Generative AI Controller so your ServiceNow instance can recognize and use it.

   For more information, see [Create a custom embedding model](https://www.servicenow.com/docs/access?context=create-byom&version=brazil&pubname=brazil-platform-administration&ft:locale=en-US).
3. Modify the prompt configuration to write your own prompts to send to the Azure OpenAI LLMs.

   For more information, see [Configure a generic large language model (LLM) connector](https://www.servicenow.com/docs/rcXZxDkgb0dTbdO5zMrroA "Connect an external LLM to the ServiceNow AI Platform by using a generic LLM connector. With a connector, you can write your own prompts to send to the LLM and create your own generative AI capabilities.").
4. Switch providers to use BYOK for AI processing.For more information, see [Manage Integration](https://www.servicenow.com/docs/04yjna9yq7OX5lqRLf_IwA "Choose the preferred integration type for configuring the available model providers. There are two ways to configure a model provider in AI Admin Hub. You can either select Original Equipment Manufacturer (OEM) or Bring Your Own Key (BYOK).").

{#byok-for-azure-open-ai__ol_r2y_zbj_yhc}
**Related tasks**   

* [Configure a custom resource path for BYOK models](https://www.servicenow.com/docs/CQYVxrRXnWa5dyP_DrNtbQ "Enter a custom resource path in your bring-your-own-key (BYOK) model configuration so that Generative AI Controller can connect to AI service providers that use a different web address than the default.")

