Aglet

Learn from query reformulation and search learning loops

A learning loop becomes useful when it preserves the path behind a reformulation, refinement, selection, or exit. Compare later sessions with the original evidence, then improve one review question about result state, intent, sample quality, or the boundary between observation and motive.

Keep the lesson for the next incident

  1. Keep the query path

    Record original phrase, result state, filters, suggestion, click, reformulation, exit, and session definition. State what the path cannot explain. This keeps an anecdotal session useful without turning it into a trend.

  2. Compare later sessions

    Review the same query family after the change and look for a new failure path or a different successful recovery. Preserve a known good and known bad session. Treat changed behavior as evidence to inspect, not proof of cause.

  3. Refine one learning check

    Require future reviews to pair a success and failure path, name the session boundary, and state the next discriminating question. Assign ownership and revisit date. Define the observation that would show the check improved search judgment rather than just reporting volume.

What to carry forward

The learning record connects one query path to later evidence and a repeatable session review. Stop when another reviewer can apply it without inferring motive from a click or exit. Carry forward uncertainty where analytics cannot identify the person’s actual intent.

Technical background: Elastic documentation.

Keep the decision with the work.

Use a Work Item in Aglet to record the problem, the evidence you have, and the next decision. Add an owner and priority, then keep updates in the discussion so the next person can follow the reasoning.

Create an account See the product workflow