AI finding severity classifier

Found something weird in your AI product? Answer four questions. We score the way we write client reports — Critical, High, or Medium.

1. What kind of issue is it?

Data leak / injection / unsafe action Wrong or hallucinated answer users will trust Broken tool call / agent step / workflow Weak fallback / confusing error state Visual / copy / minor UX polish

2. What is the worst realistic impact?

Private data exposed or irreversible action taken User makes a bad decision or loses money / trust Core journey fails; user cannot complete the task Workaround exists; support load goes up Annoyance only; no wrong outcome

3. How many users could hit it?

Default path — most users Common secondary path Rare edge case / needs intent to trigger

4. How hard is it to notice?

Silent — looks successful while wrong Confusing — user may not report it clearly Obvious failure — error or blank state

More free tools View all →

Gap finder Risk score Checklist Scope builder Prompt pack Go / no-go Triage prep

Want this checked on your real product?

Bring this output to a triage call — we'll turn it into a scoped plan.

Book a free triage All tools

GenCodeQA AI Testing. Real Quality

Independent testing for AI apps, LLM workflows, and AI-generated products.

Services
Case studies
Pricing
Resources
Guides
Blog
Tools
Method
About
FAQ
Contact