A Generative AI Engineer is deploying a conversational agent that provides financial-analysis insights to internal analysts at a company. The company has standardized on Microsoft 365 for all internal communications and collaboration. The analysts are distributed across multiple time zones and primarily work within Microsoft Teams channels, where they already discuss financial data and share reports. The agent needs to integrate with the company's existing Azure Active Directory for authentication and must respect the company's data-governance policies. Users should be able to interact with the agent within their existing workflow without switching applications.
Which interface should the engineer implement for this agent deployment?

• A. REST API endpoint with a custom frontend application hosted on the company's internal portal.
• B. Microsoft Teams integration to deploy the agent as a bot within Teams channels where analysts already collaborate.
• C. Slack integration to enable the agent in a dedicated workspace for the financial-analysis team.
• D. Databricks Apps to create a custom web application that analysts can access through their browsers with SSO authentication.

A Generative AI Engineer is deploying an agent using Mosaic AI Model Serving. The agent needs to access various Databricks resources, including Vector Search, Databricks SQL, and Functions. They need to find the easiest and best-practice way to authenticate the deployed agent to access these resources.
What approach should they choose?

• A. Log the authentication token while logging the agent; this token will be automatically used for authentication.
• B. Define resource dependencies while logging the agent and deploy it with the Agent Framework.
• C. Set appropriate permissions on the Model Serving endpoint for the agent, as these permissions will be used when connecting to other resources.
• D. Embed authentication credentials within the agent's code to access the required resources.

Databricks offers a number of built-in AI judges that provide metrics and rationale for different types of quality issues a Generative AI application may have.
Which of the following pairs of judges both require a ground-truth label in the evaluation dataset field expected_response to execute?

• A. context_sufficiency, correctness.
• B. correctness, groundedness.
• C. relevance_to_query, chunk_relevance.
• D. guideline_adherence, chunk_relevance.

A Generative AI Engineer has been asked to build an LLM-based question-answering application. The application should take into account new documents that are frequently published. The engineer wants to build this application with the least cost and least development effort and have it operate at the lowest cost possible.
Which combination of chaining components and configuration meets these requirements?

• A. For the application a prompt, a retriever, and an LLM are required. The retriever output is inserted into the prompt which is given to the LLM to generate answers.
• B. For the question-answering application, prompt engineering and an LLM are required to generate answers.
• C. The LLM needs to be frequently with the new documents in order to provide most up-to-date answers.
• D. For the application a prompt, an agent and a fine-tuned LLM are required. The agent is used by the LLM to retrieve relevant content that is inserted into the prompt which is given to the LLM to generate answers.