A prompt template needs to calculate a pet's current age in months from its stored date of birth before including that computed value in a generated response, a calculation the declarative merge-field and related-list grounding options cannot perform on their own. Which grounding technique is appropriate for this need?
Select an answer to reveal the explanation.
Short Explanation
When the grounding need involves actual computation - not just fetching a stored value - that's a job for a Flow or Apex action wired in as a grounding source, not a search technique or a simple field reference.
Full Explanation
The correct answer is an Apex invocable action or a Flow used as a grounding source, because the requirement involves computation (deriving age in months from a birth date) that declarative merge fields and related lists cannot perform on their own - Flow and Apex grounding exist precisely to bring in logic-derived values that need to be calculated rather than simply read off a field. A vector retriever performs semantic search over content for relevance matching; it has no role in numeric calculation and calling it 'a form of semantic search' misapplies the concept entirely - calculation and semantic retrieval are unrelated capabilities. A keyword search against the Patient object would retrieve or match text content, not compute a derived numeric value, so it cannot produce the needed age calculation either. The Einstein Trust Layer's audit trail is a logging and compliance feature that records prompt and response activity for review - it plays no role in performing or supplying a calculation as grounding input. The exam-relevant distinction is that Flow and Apex grounding are the technique to reach for whenever the requirement needs computed or logic-derived values rather than a value that already exists as-is on a record.