Fact Checker Agent

The Fact Checker Agent is the third major axis of measurement in Genezio, alongside Visibility and Recommendation.

Visibility asks how often you appear. Recommendation asks how often you're picked. Fact Checker asks a different question:

When answer engines talk about your brand, do they get the facts right?

You feed in specific verifiable claims about your brand — pricing, features, founding year, customer counts, certifications, anything that can be true or false — and the platform tests them against every supported answer engine.

Why Fact Checker Exists

Answer engines hallucinate. They mix outdated information with current. They confuse brands with similar ones. For most brands this is annoying. For some — security companies, financial services, regulated industries, healthcare, anyone whose product details have to be accurate — it's a real risk.

Until Fact Checker, there was no way to systematically monitor what answer engines are saying about specific facts. You'd find out about a hallucinated claim only when a customer mentioned it, or when someone happened to ask the right question.

Fact Checker turns that into measurement: you list the claims you care about, and the platform reports, per claim, per engine, whether each engine confirms it, contradicts it, or doesn't engage with it.

How It Works

A Fact Checker topic carries a set of claims to verify — claims you assert are true about your brand. For each claim, the platform runs conversations with every supported answer engine and classifies the response into one of three outcomes:

True — the engine confirms the claim
False — the engine contradicts the claim
Indecisive — the engine doesn't take a clear position either way

This three-state outcome is per-claim and per-engine, so a single claim can be confirmed by ChatGPT, contradicted by Perplexity, and ignored by Claude — and you'd see that in the platform.

Where Claims Come From

You have two ways to feed claims into a Fact Checker topic:

From the Knowledge Base

Pull claims from your brand's Knowledge Base — the documents, URLs, and snippets that already represent your source of truth. This is the natural way to operate once your KB is populated: the truth is already documented, and Fact Checker just measures whether answer engines reflect it.

Custom Claims

Type in claims directly on the Fact Checker topic. Useful for one-off measurements ("does ChatGPT know we acquired Company X last quarter?"), for testing claims that aren't yet in the KB, or for monitoring specific high-risk facts (a certification, a regulatory status, a pricing detail).

You can mix both on a single topic — some claims sourced from the KB, others typed in directly.

Where You See Results

On the Conversation

Each conversation has a dedicated Fact Checker tab showing which claims the answer engine claimed and which it didn't. Cards show a clean CLAIMED / NOT CLAIMED label so you can scan a single conversation quickly.

On the Topic and Scenario Drawers

Open the View More drawer on a topic or scenario and you'll find Fact Checker charts, including a "Claimed metric by answer engine" breakdown — which engines back your claims and which ones don't. This is the view you want when reporting up: a snapshot of whose facts are being confirmed across the engine landscape.

KPI Impact

Fact Checker conversations do not count toward AI Visibility or AI Recommendations.

The brand and the claims are named explicitly in the prompt, so the conversations don't measure organic discovery — they measure accuracy. Including them in the KPIs would distort the metrics.

Like Introspector and Comparer, Fact Checker is a measurement tool of its own: separate from visibility and recommendation, but every bit as useful for the right brand.

The Accuracy Stack

Fact Checker is the final piece of Genezio's accuracy stack:

Knowledge Base — defines what's true about your brand
Grounded Perceptions — show whether claims pulled from organic conversations match your KB
Fact Checker — actively tests whether answer engines confirm specific claims when asked

Together, these give you a complete view of accuracy:

the KB is the source of truth
Grounded Perceptions measure what answer engines spontaneously say about you
Fact Checker measures what answer engines say when asked directly

The first observes; the second probes. Read them together and you know both the narrative answer engines volunteer and the answers they give when pushed.

When to Use Fact Checker

Fact Checker is most valuable for brands where factual accuracy directly affects buyer trust or regulatory standing:

Security and financial services — where misstatements about certifications, compliance, or features can cost trust
Healthcare and regulated industries — where claims about products carry legal weight
Brands with frequently-confused competitors — where engines mix your facts with someone else's
Brands after a major change — new pricing, an acquisition, a rebrand — where stale information lingers in answer engines for months
Any brand worried about hallucinated misinformation — Fact Checker turns that worry into a measured signal

Typical Workflow

Populate the Knowledge Base (or have it ready) — the truth source you'll measure against.
Create a Fact Checker topic — attach KB-sourced claims, custom claims, or both.
Run the topic — conversations execute against every supported answer engine.
Read the results — Conversation tab for individual outcomes; topic/scenario drawer charts for cross-engine summaries.
Act on false outcomes — fix the underlying public content (your website, your docs, third-party listings) so answer engines have the right information to retrieve next time.
Re-run periodically — answer engines evolve. Yesterday's confirmed claim can become tomorrow's contradiction.