QuestionQ22

VerifiedImplement generative AI and agentic solutions

You have a Microsoft Foundry project containing an agent that your company's financial analysts use.

You need to optimize the agent workflow by adding data access and processing capabilities that meet these requirements:

  • The agent must be able to perform calculations during conversations.
  • The agent must be able to access up-to-date information from public websites.
  • The agent must be able to retrieve information from documents uploaded directly to it.

Which tool should you use for each requirement? A tool can be used once, more than once, or not at all.

Drag & Drop
Access up-to-date information from public websites:
Perform calculations during conversations:
Retrieve information from documents uploaded directly to the agent:
Explanation

Grounding with Bing Search lets an agent retrieve current, real-time information from the public web, which is required to answer with up-to-date information from public websites. Code interpreter runs code in a sandboxed execution environment, letting the agent perform calculations and data processing during a conversation. File search indexes and retrieves content from documents uploaded directly to the agent, satisfying the third requirement. Computer use and Microsoft Fabric are not designed for these scenarios: computer use drives a GUI/OS environment, and Microsoft Fabric integration targets enterprise data platform assets rather than files uploaded directly to the agent.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!