- Home
- Skills
- Research & Analysis
- AI Brand Fact Checker
Works with the AI Agents you already use
AI Brand Fact Checker
Audit AI assistant answers against your source of truth to catch hallucinations and incorrect brand claims.
$9
AI Brand Fact Checker
Example session with this skill installed
Verify these ChatGPT and Perplexity answers against our official pricing.md. Tell me if they got the $29/mo starter price and our refund policy right.
- Read your context and instructions
- Compiled the ai brand fact
I have completed the fact-check of the provided AI answers against pricing.md for Northwind CRM. The audit checked 42 individual claims across two engines.
AI Brand Fact Check Report: Northwind CRM
| Claim ID | Status | AI Claim | Evidence / Source Quote |
|---|---|---|---|
| 1 | CONTRADICTED | "The Starter plan costs $49 per user per month." | source says $29; "Our Starter plan is priced at $29 per user, billed monthly." |
| 2 | EVIDENCED | "Northwind offers a 30-day money-back guarantee." | source says: "We provide a full refund within the first 30 days of service." |
| 3 | NO_EVIDENCE | "Includes 24/7 phone support for all tiers." | not in source; The provided documentation does not mention phone support hours. |
| 4 | CONFLICT | "Storage limit is 10GB (ChatGPT) / 5GB (Perplexity)." | source says 5GB; "Standard storage is capped at 5GB for all non-enterprise users." |
3. Cross-engine conflicts
- Pricing: ChatGPT claimed $49/mo while Perplexity claimed $29/mo. The source confirms $29/mo.
- Storage: Engines disagree on storage limits (10GB vs 5GB). The source confirms 5GB.
4. Documentation Gaps
- The source does not specify support hours, leading ChatGPT to claim 24/7 availability.
- There is no mention of a "Free Tier," which both engines attempted to describe.
Next steps
- Update the public pricing page to emphasize the $29 price point to overwrite the $49 hallucination.
- Add an explicit "Support Hours" section to your FAQ to prevent AI from promising 24/7 phone access.
- Explicitly state "No Free Tier available" in your product brief to trigger
CONTRADICTEDflags in future audits.
Connects securely to your tools. The creator never sees your data.
About this skill
AI assistants frequently hallucinate brand facts, pricing, and policies, leading to customer misinformation. This skill identifies exactly where ChatGPT, Gemini, and Perplexity deviate from your official documentation by auditing AI responses against your source of truth.
What it does
- Evidence extraction pulls verbatim passages, names, and figures from your source files to verify or debunk AI claims.
- Contradiction detection flags claims that are explicitly denied by your documentation, such as non-existent features or tiers.
- Conflict analysis identifies discrepancies where different AI engines provide different values for the same attribute, like pricing.
- Gap identification lists claims that your source material does not cover, highlighting where your documentation needs more detail.
- Fix recommendations suggests specific updates to your public-facing content to prevent future AI hallucinations.
How it works
- Collect AI answers by pasting responses from various engines into a markdown file, separated by headers or dividers.
- Provide a source of truth such as a pricing page, product brief, or FAQ file in text or markdown format.
- Execute the check using the
claim_check.pyscript to compare the answers against the source and generate a report. - Review the evidence produced by the tool, which categorizes findings as contradicted, conflicted, or unsupported.
Frameworks & tools
This skill uses a Python-based analysis script (Python 3.9+) that requires only the standard library. It operates entirely offline without external API calls or network access.
Why this beats prompting it yourself
Manual fact-checking of long AI responses against dense documentation is prone to human oversight. This skill automates the word-for-word and figure-for-figure matching, ensuring every claim is accounted for and backed by a quoted source sentence.
Use cases
- Auditing AI search results to ensure product pricing is reported accurately.
- Comparing how different LLMs describe your company's service level agreements.
- Identifying missing information in your documentation that leads to AI hallucinations.
- Verifying that deprecated features are no longer being claimed by AI assistants.
Known limitations
The tool matches text and figures rather than judging abstract "truth." It cannot verify if links are active as it has no network access. Final semantic judgment remains with the user.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Research & Analysis.
Requires manual collection of AI answers. Does not judge truth, only matches evidence. No network access to verify links. Semantic judgment requires human review.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 27 days ago
- Software
Python