A credit team wants a score estimating likelihood a case needs specialist review, based on historical Dataverse features, displayed in the agent before escalation. Which capability fits better than generative answers alone?
Select an answer to reveal the explanation.
Short Explanation
You need prediction models trained on labeled history to return risk scores. Generative answers summarize documents, so don't use FAQ bots when you need trained scores.
Full Explanation
AI Builder prediction models address structured ML scoring; generative answers address language question answering over content. The prediction model integrated through Power Automate/Dataverse is correct because it uses historical features and returns the risk score to the topic for branching, unlike FAQ answering. Marketing-site generative answers, static 42 message, and disabling Dataverse security roles do not produce valid predictive score. Exam tip: Classification/prediction → AI Builder; document question answering → generative answers + knowledge.