A practical starting point
Work through these checks in order. Use the results to decide what needs to change in your own workflow.
- Collect retrieval cases where relevant evidence appears below irrelevant candidates.
- Evaluate a reranker using known relevance judgments.
- Compare final answer support before adopting it broadly.
A large candidate pool may contain the correct policy but need better ordering before context selection.
A mistake to avoid
Do not add another model stage without evidence that ranking is the bottleneck.
How to check progress
Track relevance at the selected cutoff and end-to-end response time.
Your action checklist
Use this checklist while reviewing your own setup. Checkmarks are temporary and stay in this open page.
Need the wider workflow?
This quick note covers one decision. For a worked example, decision table and a fuller implementation workflow, read When retrieval is the right addition to an AI assistant.
Further reading
Official background resources for this topic. The workflow above is general guidance, not a guarantee of a particular result.
AI-assisted editorial content. Examples are illustrative, not client case studies. Read our editorial policy or report a correction.