Run a structured fact-check before the final answer

Verification workflow for checking claims, citations, evidence boundaries, Chain-of-Verification, and confidence before using AI-generated output.

Purpose

This guide defines a final fact-checking and claim-verification workflow. It is not the evidence-boundary chooser. First select or confirm the active evidence boundary, then use this workflow to decompose the draft or answer into claims, verify each claim against the allowed evidence, check citation or artifact locators, run Chain-of-Verification, and return either a corrected final answer or a fail-closed result.

When to use this

Use this section to decide whether this workflow is the right fit before you configure prompts, policies, or reference material.

Use case
A generated answer contains factual claims
Use this when an AI-generated answer, draft, report, summary, article, or technical explanation contains claims that need evidence support before final use.
Use case
The answer will be published, sent, reused, or used for a decision
Use this when unsupported claims, weak citations, or invented evidence would create publication, product, academic, engineering, operational, legal, or reputational risk.
Use case
The workflow already has an active evidence boundary
Use this after selecting files-only, cited authoritative sources, web verification, or academic evidence mode. The fact-checking workflow verifies within that boundary; it does not replace the boundary decision.
Use case
The output needs a final confidence signal
Use this when a non-sentinel final answer should include an evidence-based confidence score after claim support, citation quality, and uncertainty have been checked.

Workflow assets

Required workflow assets

Open the prompts, policies, and reference pages needed to run this workflow correctly.

Required prompt
No simulation / no fabrication
Blocks invented facts, fabricated sources, simulated scans, guessed evidence, and unsupported system behavior during verification.
Type: Guardrail component
Belongs in: Instruction or verification layer
Use when: The workflow must prevent invented evidence or simulated verification.
Open prompt
Required prompt
Do Not Confirm Without Evidence
Prevents the assistant from confirming user-provided claims or draft claims without evidence support.
Type: Evidence-control component
Belongs in: Verification layer
Use when: The draft contains claims that may be plausible but not yet verified.
Open prompt
Optional prompt
Analyze before answering
Adds a short evidence and constraint review before the final answer is produced.
Type: Review-preparation component
Belongs in: Runtime prompt layer
Use when: The answer is complex, multi-source, or likely to contain unsupported conclusions.
Open prompt
Optional prompt
Confidence score
Adds evidence-based confidence reporting after claim support, citation quality, and uncertainty are checked.
Type: Verification reporting component
Belongs in: Verification layer
Use when: A non-sentinel final answer needs explicit confidence reporting.
Open prompt
Optional policy
Use only the files you provide for factual answers
Use when the active evidence boundary is provided artifacts only.
Type: Evidence-boundary policy
Controls: Artifact-only claim support, file-location locators, and fail-closed missing-artifact behavior.
Read policy
Optional policy
Use cited public sources
Use when the active evidence boundary is cited authoritative public sources.
Type: Evidence-boundary policy
Controls: Authoritative-source requirements, stable locators, citation discipline, and fail-closed missing-source behavior.
Read policy
Optional policy
Use web search and cite sources
Use when the active evidence boundary allows browsing or public-source search for current or niche claims.
Type: Web-verification policy
Controls: Browsing availability, citation requirements, source quality, and insufficient-evidence handling.
Read policy
Optional policy
Add an evidence-based confidence score
Use when successful non-sentinel responses must end with an evidence-based confidence line.
Type: Confidence reporting policy
Controls: Confidence reporting after source support, coverage, and uncertainty are checked.
Read policy
Required reference
Prompt layers and policy mapping
Explains where instruction rules, source material, runtime prompts, and verification gates belong.
Type: Configuration reference
Use when: Use when configuring final verification in ChatGPT, Claude, Gemini, or an internal AI system.
Open reference
Optional reference
Verification techniques
Use when you need the broader verification taxonomy behind claim checking and Chain-of-Verification.
Type: Verification reference
Use when: Use when selecting a verification pattern or explaining why final verification is required.
Open reference

Implementation procedure

Step-by-step implementation procedure

Follow the workflow in order. Each step gives one action and one verification check before continuing.

  1. Runtime prompt layer

    Confirm the active evidence boundary

    Identify whether the answer must be checked against provided artifacts, cited authoritative sources, web-verified sources, or academic evidence mode.

    Action
    If no evidence boundary is selected, stop and route to the evidence-boundary decision guide before fact-checking.
    Verify
    The review states the active evidence boundary before evaluating claims.

  2. Runtime prompt layer

    Separate the draft into checkable claims

    Extract material factual claims, technical claims, source claims, quantitative claims, claims about standards or documentation, and claims that affect the conclusion.

    Action
    Create a claim list and exclude purely stylistic or subjective wording unless it implies a factual claim.
    Verify
    Each material claim can be matched to evidence, corrected, removed, or marked unsupported.

  3. Reference layer

    Map each claim to allowed evidence

    Check every material claim against the source type allowed by the active boundary.

    Action
    Use artifact locators for files-only claims, stable citations for public-source claims, and reviewed-source locators for academic evidence mode.
    Verify
    No claim is treated as verified because it sounds plausible or because it appeared in memory or prior chat context.

  4. Verification layer

    Check citation and locator quality

    Verify that citations, file references, section references, source names, and locators actually support the claims they are attached to.

    Action
    Flag weak, vague, mismatched, stale, inaccessible, or non-authoritative citations according to the active policy.
    Verify
    Every retained claim has a locator that supports the exact claim, not just the general topic.

  5. Verification layer

    Run Chain-of-Verification

    Run a final second-pass check for unsupported claims, contradictions, overstatement, missing evidence, and boundary violations.

    Action
    Correct supported wording, remove unsupported claims, or return the fail-closed result required by the active policy.
    Verify
    The final answer is aligned with the evidence actually available.

  6. Verification layer

    Apply confidence reporting only when allowed

    Add confidence only after verification passes and only when the active policy does not require a sentinel-only fail-closed output.

    Action
    If evidence is insufficient, do not append a confidence score to a sentinel-only response. If the answer is supported, report confidence based on evidence quality, coverage, and uncertainty.
    Verify
    Confidence does not replace citations, artifact locators, or missing-evidence handling.

  7. Runtime prompt layer

    Return the final verified output

    Return a corrected final answer, a claim-level correction report, or the fail-closed result required by the active evidence boundary.

    Action
    Keep the output format aligned to the user’s requested deliverable and the active policy.
    Verify
    The final answer does not include unsupported claims that were identified during verification.

Verification checklist

Use this checklist before accepting the output, publishing it, or using it as evidence for a downstream workflow.

Boundary
The active evidence boundary is named
The review identifies whether claims are checked against provided artifacts, cited authoritative sources, web-verified sources, or academic evidence mode.
Claims
Material claims were extracted
The workflow identifies claims that affect factual accuracy, technical correctness, source support, or the final conclusion.
Evidence
Each retained claim has allowed evidence
Every retained factual claim is supported by an allowed artifact locator, cited source, reviewed source, or other evidence permitted by the active policy.
Citations
Citations and locators support the exact claim
The cited source or artifact locator supports the claim attached to it, not merely the surrounding topic.
Fail closed
Unsupported claims stop or change the output
Unsupported claims are removed, corrected, marked as insufficient evidence, or handled with the active fail-closed sentinel.
Confidence
Confidence is evidence-based and non-sentinel only
A confidence score is added only after verification passes and never appended to a sentinel-only fail-closed response.

Next step