AI Workflows Library · Practical AI use cases & templates
EN RU

Quality checklist (anti‑hallucination)

Use this before you copy any AI output into “real life”.

10 checks

  1. Facts: any numbers, dates, names — are they real?
  2. Assumptions: are assumptions explicit?
  3. Scope: does it answer the question you asked (not another one)?
  4. Missing pieces: what info was required but absent?
  5. Contradictions: do parts conflict with each other?
  6. Actionability: can someone execute it step by step?
  7. Tone: matches the audience (neutral/friendly/strict)?
  8. Risk: any legal/medical/financial claims → remove or cite sources.
  9. Safety: no harmful instructions.
  10. Next step: one clear next action.

Copy‑paste “QA pass” prompt

Quality audit this text.
Mark issues under: Facts, Assumptions, Missing, Contradictions, Risk.
If a fact is uncertain, label it UNKNOWN.
Then propose a corrected version.
Text:
<<<TEXT>>>

Next