Prompt

Software EngineeringPromptIn a pack

Search Relevance Tuning — Judgement Lists, Offline Metrics, Boosting and Safe Experiments

Improves search result quality with evidence — collecting top and zero-result queries, building judgement lists, measuring with offline metrics, tuning field boosts, function scores and business rules, and running controlled experiments — instead of tweaking boosts until one query looks right.

  • Role-Based
  • Step-by-Step
  • Constraints
  • Structured Output

Opening lines · ~342 words in full

ROLE: You are a relevance engineer. Changing a boost to fix one complaint often breaks dozens of other queries, so you measure relevance on a representative set of queries, change one thing at a time and confirm improvements with real users.

CONTEXT: The search experience and what a good result means …

Written by the PromptCorrectly team© 2026 PromptCorrectly · All rights reserved

How to use it

  1. Read it, then replace anything in [BRACKETS] with your details — the more concrete the context, the sharper the answer.
  2. Copy it (or download the .md) and paste it into the AI you already use — it knows your work, so that is where the prompt does the most.
  3. Want a sharper answer? Ask the AI to check its own draft against your brief before it replies — that is self-critique prompting.

Techniques in this prompt

Role-Based

Assigns the model an expert persona so it adopts the right vocabulary, depth, and standards for the task.

Learn this technique
Step-by-Step

Forces explicit intermediate reasoning instead of jumping to a conclusion, which improves accuracy on hard tasks.

Learn this technique
Constraints

Sets the rules and boundaries — tone, length, what to avoid — that keep the output on-target.

Learn this technique
Structured Output

Pins the response to a defined structure so it drops straight into your workflow.

Learn this technique

Works with

Any chat AI — ChatGPT, Claude, Gemini, Copilot, Grok, Mistral or a local model. The structure does the work, so you are not tied to one vendor or one model version.

New to structured prompts? Start with how to prompt AI, the RCTCO prompt framework this prompt is built on, and role prompting examples.

More in Software Engineering

Production Incident Root Cause AnalysisDrives a disciplined RCA from symptoms to root cause and prevention, separating contributing factors from the true trigger.Security-Focused Code Review For Pull RequestsReviews a diff specifically for security vulnerabilities, mapping findings to severity, exploit path, and concrete fixes.Legacy Code Refactoring StrategistPlans a safe, incremental refactor of tangled legacy code with characterization tests and reversible seams.API Contract Designer With OpenAPI OutputDesigns a consistent, versioned REST resource and emits a ready-to-use OpenAPI 3.1 fragment plus error model.