Truth Check

AI brand monitoring that checks whether AI gets you right

Being mentioned is only half of AI brand monitoring. The other half is whether the mention is true. Truth Check reads what ChatGPT and the other AI engines you track say about your brand, pulls out each factual claim, and checks it against your own facts: pricing, features, availability, integrations, policies. Wrong claims come with the correction, the evidence and a drafted fix.

Free Trial See the platform

7-day trial · Cancel anytime

A WriteWorks Truth Check claim marked Wrong: AI engines say Surge45 charges a fixed £2,500 a month while Surge45's own sources say retainers start at £4,000 a month, with the quoted evidence and three ways to fix it (sample data).A WriteWorks Truth Check claim marked Wrong: AI engines say Surge45 charges a fixed £2,500 a month while Surge45's own sources say retainers start at £4,000 a month, with the quoted evidence and three ways to fix it (sample data).

Connects to the tools your team already uses

The platform

See WriteWorks in Action

This is the working product, not a mockup: the same visibility scores, briefs and published content your team would be looking at. Start a free trial and point it at your own domain, or take the tour of what sits behind the screen.

Get started
Surge45
Home
Ask AI
Opportunities128
Research
Create
Measure
AI visibility
AI discoverabilityNew
AI crawlers
Reports
Context
Support
Settings
Sign out
SCSam Cartersam@example.com
WRITEWORKS
Measure/AI visibility
Search brands, prompts, competitors…⌘K
New
1.2M (62%) leftAsk AI
The challenge

Why mentions aren't enough

An AI answer can name you and still get you wrong.

AI repeats old or wrong facts with confidence

A retired price, a discontinued plan, an integration you never built. Buyers read it in an AI answer and take it as fact.

Mention tracking can't tell true from false

Counting brand mentions in ChatGPT shows you are named. It doesn't show that the answer gets your pricing wrong.

Nobody knows where the error came from

Without the sources behind an answer, there's no way to tell whether the wrong fact came from your own outdated page, a competitor or a third-party review.

One mistake, many engines

The same wrong claim shows up on several engines and dozens of prompts, and gets treated as dozens of separate problems.

The solution

How Truth Check works

Extract the claims, check them against your facts, trace the source, fix it.

Truth Check extracts the factual claims about your brand from each new answer, skipping opinions, recommendations and claims about other companies. Each one is checked against your Context: records, team answers, brand profile, documents and synced pages. The checker is instructed not to use its own knowledge.

◆How Truth Check worksLIVE
WriteWorks Truth Check: a wrong claim and how to fix it. Close-up: AI says vs Your sources say, with the evidence (sample data).WriteWorks Truth Check: a wrong claim and how to fix it. Close-up: AI says vs Your sources say, with the evidence (sample data).
Success story

What real brands fixed with WriteWorks

What each brand fixed, measured before and after with the same audit, in their own words.

Crawler responses 2.5x faster

“Turning caching on made our pages 2.5 times faster for crawlers. While we were in there, we fixed the headings too: guides with a clean heading structure went from 32% to 100%.”

S4
Surge45 teamRead the Surge45 success story
Capabilities

What you can see

Accuracy by engine, theme and week.

Accuracy by engine

See which AI engines get your facts right most often.

Accuracy by theme

Pricing, Features, Availability, Integrations, Company, Policies, Compliance, Support and Other.

Weekly accuracy trend

Track whether corrections are taking hold.

Fact check on any answer

A fact-check block alongside stored engine answers.

Gaps become questions

A Can't verify claim becomes a question in Context; answer it once and future checks use it.

Fix workflow

Mark fixing, resolve, dismiss, or flag a wrong verdict.

Time savings

What changes

From asking ChatGPT yourself to a weekly fact check.

TaskBeforeWith WriteWorksTime saved
Find out what ChatGPT says about your pricingAsk it yourself, prompt by promptEvery tracked prompt checked weeklyThe manual sweep
Prove an answer is wrongHunt for the sourceCorrection plus evidence from your ContextResearch
Trace a wrong fact to its sourceGuessCited pages labelled own, competitor or third partyDetective work
Fix itWrite from scratchDrafted page copy, brief or outreach requestPer claim
Included

What's included

AI brand monitoring with evidence.

Factual claims extracted from AI answers about your brand
Checked against your Context: records, team answers, profile, documents and synced pages
Verdicts: Accurate, Wrong, Can't verify
Corrections with quoted evidence for every wrong claim
Cited sources labelled own page, competitor or third party
Similar wrong claims grouped across engines
Drafted fixes: page copy, content brief or outreach request
Weekly checks per prompt per engine, plus Check now
Accuracy by engine, by theme and over time
Uses AI tokens from your plan
Built for

Who uses it

Anyone responsible for what buyers believe about the brand.

Product marketers

Catch wrong positioning, features and pricing before buyers repeat them.

Brand and comms teams

AI brand monitoring with evidence, not just a mention count.

Legal and compliance

Spot inaccurate claims about policies, compliance and availability.

Agencies

Show clients exactly what AI gets wrong about them, and fix it.

Featured story

“A near-perfect audit score made it easy to think the job was done. WriteWorks found the one gap that mattered: none of our homepages said who we are in a way a machine could read. We added Organization markup in every market and dated every FAQ page, and key pages carrying the schema AI expects went from 92% to 100%. Small changes, and a much clearer picture of GraftPal for AI.”

GraftPal team

Read the full story
0% → 100%
Market homepages with Organization markup
20% → 100%
Articles and FAQs with a modified date
6.7x
Faster median server response to crawlers
FAQ

Frequently asked questions

Everything teams typically ask before getting started.

What is AI brand monitoring?+
AI brand monitoring is tracking what AI assistants and answer engines say about your brand. It has two parts: whether you are mentioned at all, and whether what is said is right. WriteWorks covers the first with mention tracking and sentiment, and the second with Truth Check, which fact-checks the claims inside each answer.
How do I see brand mentions in ChatGPT?+
Track the prompts your buyers ask. WriteWorks runs them on ChatGPT and the other engines on your plan, stores every answer and shows where your brand is named. Truth Check then reads those answers and checks each claim about your brand against your own facts.
Is Truth Check an AI fact checker?+
For claims about your own brand, yes. It isn't a general fact checker: it checks only factual claims about your company, such as pricing, features, availability, integrations and policies, and only against the facts you hold in Context. It doesn't use the model's own knowledge to judge them.
Where do the facts come from?+
Your Context: records, team notes and answered questions, your brand profile, documents and synced pages. Team notes and answered questions take priority over documents. If a claim can't be confirmed either way it's marked Can't verify and becomes a question for your team.
How often does it run?+
Once a week for each tracked prompt on each engine, in small batches through the day, plus a Check now button. The weekly accuracy trend needs two weeks of data.
Does it use AI tokens?+
Yes. Checking claims and drafting fixes both use AI tokens from your plan, within a monthly allowance for Truth Check so it can't use up the whole plan.
How is Truth Check different from mention tracking?+
Mention tracking tells you whether and how often AI engines name your brand, and alerts you when that changes. Truth Check tells you whether what they say is accurate. Sentiment sits in between, scoring the tone of each mention.
Mentioned, described, accurate

Three questions about every AI answer

Are you mentioned?

AI mention tracking: how often engines name you, and alerts when that moves.

See mention tracking

How are you described?

Sentiment: the tone of each mention, yours and your competitors'.

See sentiment

Is it true?

Truth Check: each factual claim checked against your own facts, with the fix.

Explore more

Keep exploring

Related solutions, adjacent use cases, and platform features.

Pillar

Related AI visibility features

Fix the facts

Find out what AI gets wrong about your brand

Truth Check fact-checks every claim AI engines make about your brand, traces where it came from and drafts the fix.

Start free trial See the platform