HIL-PER-01Tier IIHuman-in-the-Loop Checkpoints

User-Intent Clarification

When a user's request is ambiguous or underspecified, the agent asks a focused clarifying question before acting — rather than guessing.

Runtime: PerceiveAgentOps: Operate & MonitorSee on the matrix →
USER-INTENT CLARIFICATION User Request Ambiguity Detectorsignals ·confidence Clear · Proceedconfidence high Ambiguous · Askone focused question Missing Param · Askthe specific fieldOne clarifying question, not a quiz. Cache disambiguation for the session.
In Plain English

This page is the complete instruction page for one pattern called "User-Intent Clarification." It explains the problem this pattern solves, the idea behind the solution, when you should (and should not) use it, and what happens afterward — both the good effects and the costs.

This matters because building AI agents is not just about making them clever. It is also about making them safe and predictable. Following a well-tested pattern like this one helps avoid common mistakes, and shows you exactly which safety rules and regulations it connects to, listed under "Standards Mesh" on this page.

Context

Users phrase requests imprecisely. An agent that guesses confidently produces plausible-but-wrong outputs. Clarification is cheap; recovery from wrong action is expensive.

Problem

Guessing intent on ambiguous requests silently transfers risk from the user to the agent.

Forces

  • Interaction friction vs accuracy
  • Question quality vs user patience

Solution

Detect ambiguity via uncertainty signals: multiple plausible interpretations, missing required parameters, confidence below threshold. Pose a single, focused clarifying question rather than a multiple-choice quiz. Cache disambiguation for the session.

Applicability

  • Conversational agents
  • Task-automation agents with side effects
  • Any agent where misinterpretation has real consequences

Anti-Patterns

  • Silently picking the most common interpretation
  • Asking several clarifying questions at once

Consequences

  • +Higher task-completion accuracy
  • +Users feel heard
  • Marginal latency per clarifying turn
  • Over-asking annoys