← Back to Dashboard
1. Offline Evaluation Design2. Online Eval and User Feedback Loops

Online Eval and User Feedback Loops

📚 Evaluation Frameworks10 min90 XP

Production Learning Loops

Offline evals prove readiness; online evals prove reality. Once real users arrive, their behavior becomes your richest - and cheapest - evaluation signal.

Signal Sources, Strongest First

SignalWhat it tells youCaveat
Task completionDid the user accomplish the goal (ticket resolved, doc accepted)?Define "completion" per workflow explicitly
Behavioral signalsCopy/paste of answer, retry with rephrasing, session abandonmentImplicit - correlate, don't over-interpret single events
EscalationsHuman takeover rate and reasonsRising escalations = quality or trust regression
Thumbs up/downDirect sentiment on specific responsesLow response rate, negativity bias - use as trend, not truth
Incident outcomesPostmortems attributing impact to AI behaviorLagging but highest-severity signal

Closing the Loop

  1. Sample and label - pull flagged interactions (bad feedback, escalations, abandonments) into a review queue.
  2. Diagnose the layer - was it retrieval, prompt, routing, or guardrail? (Your trace IDs make this a query, not an investigation.)
  3. Feed the datasets - every confirmed failure becomes a regression-set case; every novel attack joins the adversarial set.
  4. Adjust the lever - prompt revision, routing threshold, guardrail tuning - then verify offline before redeploying.

This is a flywheel: production failures continuously harden the offline suite, which prevents their recurrence.

Anti-pattern: dashboards nobody acts on. Every online signal needs a routing rule - who reviews it, within what SLA, feeding which dataset. Telemetry without a consumer is decoration.
🧪 Knowledge Check
Press 1-4 to select1 of 2
What distinguishes online evals from offline evals?
They use no metrics
They include real user behavior and outcomes
They cannot run in production
They ignore safety