KP.
← All posts
AI Quality · 25 Jun 2026 · 1 min read

Where AI Fits Into a Modern QE Strategy

AI won't replace your test strategy — but LLM-as-judge, change-scoped impact analysis, and good guardrails will reshape it. Where to start.

Everyone wants "AI in QE." Fewer people can say where it actually pays off. Here's the map I use.

Three places AI earns its seat

  1. LLM-as-judge for evaluating output correctness where assertions are fuzzy
  2. Change-scoped impact analysis to run the right regression, not all of it
  3. Draft generation for tests and fixtures — always human-reviewed

Guardrails first

Prompt libraries, human-in-the-loop review rubrics, and data-handling rules aren't overhead — they're what make AI-assisted work safe to ship.

Start small, measure, and keep a human in the loop.

← All postsGet in touch →