Universal Containers (UC) has implemented Generative AI within Salesforce to enable summarization of a custom object called Guest. Users have reported mismatches in the generated information. In refin

Question

Options

• AEnable prompt test mode, allocate different prompt variations to a subset of users for evaluation,
• BCreate concise, clear, and consistent prompt templates with effective grounding, contextual role-
• CSubmit a prompt review case to Salesforce and conduct thorough testing In the playground to

• A
  12% (6)
• B
  82% (40)
• C
  6% (3)

Explanation

When AI-generated summaries contain mismatched or inaccurate information, the root cause is almost always poor prompt design - specifically: vague instructions, insufficient grounding, or lack of contextual role guidance. The correct fix is to create concise, clear, and consistent prompt templates that use effective grounding (linking the prompt to authoritative Guest object data) and contextual role-based framing so the LLM understands what it should summarize and how. Option A (A/B testing with a user subset) is a validation methodology, not a design fix. Option C (submitting a case to Salesforce) is not a valid prompt engineering strategy and would not resolve a design-level issue.

No community discussion yet for this question.