← Amazon Connect Intelligence
Amazon ConnectEnhancement

Automated checks flag evaluation questions AI cannot score reliably

Managers can run a one-click check on performance evaluation forms that identifies questions lacking context for AI scoring and suggests more specific wording.

What changed

AWS added an automated check to Amazon Connect Customer evaluation form setup. It compares a form against best practices and lists questions that are missing the context AI needs to score them reliably, with suggestions to make them more specific, for example defining exactly what a required sales disclosure must contain.

Vendor documentation (opens in a new tab)

Why it matters

AI-filled evaluations are only as consistent as the questions behind them. Vague criteria produce scores that are hard to defend to agents, so a pre-flight check is a practical calibration step before scaling automated scoring to human or AI agents.

CxCave practical context

Who should care

  • Quality teams
  • Supervisors
  • Administrators

Action

REVIEW

Run the check on forms already used for AI-filled evaluations, tighten flagged questions, and recalibrate against a sample of manually scored contacts before relying on the results.

CxCave recommendation, not a vendor mandate unless explicitly stated.