Back to components

AI Interfaces

Confidence Escalation Panel

A decision boundary that converts uncertainty signals into request-context or human-review actions.

confidenceescalationreview

Installation

Copy and paste the code into your project.

Accessibility notes

Every signal has an exact text value and the escalation cause is stated explicitly rather than encoded by color.

Decision boundary

Review before answering

72% confidence
Source agreement
84%
Policy clarity
62%
Answer stability
71%

Escalation triggered

Policy clarity is below the 70% automatic-answer threshold.

Preview accent

TypeScript
const signals = [{ label: "Source agreement", value: 84 }, { label: "Policy clarity", value: 62 }, { label: "Answer stability", value: 71 }];

export function ConfidenceEscalationPanel() {
  return <section className="w-full max-w-sm rounded-xl border border-white/12 bg-[#0b0f14]/92 p-4 shadow-2xl">
    <header><p className="text-xs font-semibold uppercase tracking-[0.18em] text-[#fb923c]">Decision boundary</p><h3 className="mt-1 text-base font-bold text-white">Review before answering</h3></header>
    <dl className="mt-4 space-y-3">{signals.map((signal) => <div key={signal.label}><div className="flex justify-between text-[0.65rem]"><dt>{signal.label}</dt><dd>{signal.value}%</dd></div><div className="mt-1.5 h-1.5 rounded-full bg-white/[0.06]"><span className="block h-full rounded-full bg-[#fb923c]" style={{ width: signal.value + "%" }} /></div></div>)}</dl>
    <div className="mt-4 rounded-lg border border-[#fb923c]/20 bg-[#fb923c]/[0.06] p-3"><p className="text-xs font-semibold text-[#fdba74]">Escalation triggered</p><p className="mt-1 text-[0.65rem] text-slate-400">Policy clarity is below the 70% automatic-answer threshold.</p></div>
  </section>;
}
Usage
<ConfidenceEscalationPanel />

Related components

Answer evidence

3 claims supported

Verified
  • Release date

    Launch brief

    98%
  • Plan limits

    Pricing notes

    91%
  • Migration path

    Support guide

    86%

AI Interfaces

Answer Evidence Card

Intermediate

A claim-level evidence summary that connects generated answers to sources and confidence scores.

evidencecitationsconfidence
View component

Approval

Publish launch prompt

pending
DesignLegalOps

AI Interfaces

Approval Action Panel

Intermediate

A review panel with assignee chips and paired approval actions for launch workflows.

approvalreviewworkflow
View component

Confidence

ready
Grounded91%
Coverage84%
Risk12%

AI Interfaces

Confidence Meter

Intermediate

A multi-signal confidence card for grounded AI answers, eval reviews, and release checks.

confidenceevalsignals
View component