Library › Research AI › HR and recruiting

Pick salary surveys that fit your benchmarking rules

Practice: Setting limits and rules. About 5 minutes.

About this exercise

You're benchmarking pay for your accounting team in Phoenix. Your rules: 2026 data, Arizona or Phoenix metro, at least 100 companies, and includes the job family "Accounting & Finance". You want exactly 2 numbered picks, each with its name and a reason under 12 words, using only the survey list, plain text, nothing else.

Common mistakes when you ask AI for this

  • Not giving every rule. The AI favors the biggest or most famous survey, even if it's national or old. Try: Say: "Must meet all four: 2026, Arizona or Phoenix, 100+ companies, Accounting & Finance."
  • Inventing pay data. The AI may add salary ranges it thinks the surveys contain. Try: Add: "Use only details in my list."
  • Not saying how many. The AI decides the length itself and usually returns more picks than you can use. Try: Be exact: "exactly 2 numbered picks."
  • Getting markdown back. AI defaults to bold labels and headings, which show up as stray symbols in a note to your comp lead. Try: Ask for "plain text, no bold, no headings."

What a strong prompt includes

  • Lists the rules
  • Says exactly how many
  • Limits each reason
  • Says to use only your sources
  • Asks for plain text

Practice this now ↑

Related exercises