Library › Research & AI Search › Managers and team leads

Label a deep-research report's findings by confidence

  • For Managers and team leads
  • You practice Catching mistakes and made-up facts
  • Time about 5 minutes

About this exercise

A deep-research tool gave you 4 findings about compressed work weeks before your team vote. Label each High, Medium or Low using your rules: High = 2 or more sources from 2024 or later; Medium = exactly 1 source from 2024 or later; Low = only older sources or a source paid for by a vendor. Write exactly 4 numbered lines, each with the label and the source count and years that justify it. Plain text, under 100 words.

Common mistakes when you ask AI for this

  • Trusting the report's summary. The AI repeats "strong evidence" because the report says so. Try: Say: "Label each finding yourself using my rules, not the report's summary."
  • Not giving your rules. The AI judges confidence by gut feel, so a vendor-paid study can rate High. Try: Add: "High = 2+ sources from 2024 on; Medium = 1; Low = older or vendor-paid."
  • Labels with no proof. The AI gives a label but not the count and years, so you can't check it. Try: Add: "Each line shows the source count and years."
  • Long or formatted output. AI writes a section per finding with bold headings. Try: Add: "Exactly 4 numbered lines, each under 20 words, plain text."

What a strong prompt includes

  • Gives your confidence rules
  • Flags vendor-paid sources
  • Asks for the proof on each line
  • Sets the shape
  • Asks for plain text

Practice this now ↑

Related exercises