Configuring Work Item compose agent

You can configure the Work Item compose agent to customize the AI model and customize the default compose prompts for existing prompt set.

About this task

You can specify or update the Large Language Model that agent uses to compose work items.

You can configure the following settings for this agent:
  • AI model identifier: Specify or update the Large Language Model that agent uses to compose draft for work items.
  • Manage prompt mapping: Manage default prompt text used by the Work Item compose agent.

Procedure

  1. From the Agent management page, click the Settings icon in the Actions column for the Work Item compose agent.
  2. Optional: Turn on Show default value to view default settings.

    This toggle displays the default values for all configuration fields, helping you understand the default settings before making changes.

  3. In the Model ID field, review or update the AI model identifier.

    The default model is mistralai/mistral-small-3-1-24b-instruct-2503. For more information, see Model configuration.

    Note: The model ID format depends on your configured LLM provider:
    • For IBM WatsonX, use WatsonX model names (for example, mistralai/mistral-small-3-1-24b-instruct-2503 or meta-llama/llama-4-maverick-17b-128e-instruct-fp8)
    • For AWS Bedrock, use AWS Bedrock model names or cross-region inference profile IDs (for example, mistral.mistral-small-2402-v1:0 or us.meta.llama4-maverick-17b-instruct-v1:0)

    For more information about configuring your LLM provider, see Setting up LLM providers.