Verification
Every claim AI makes about you, checked.
We pull out what AI says about your brand and look it up on the live web, official sources first. Only an official source can mark a claim wrong.
Claims about Fernvale Skincare
Sample data“Returns are free within 14 days”
WrongOfficial source: the returns page says 30 days.
“All products are vegan”
CorrectOfficial source: the ingredients page agrees.
“Ships within two days everywhere”
Couldn’t verifyNo official page found.
↑ can’t find it is not the same as wrong
Step one · Find the claims
Pull out the facts AI states.
We go through the answers that name your brand and pull out each factual statement about it. The same claim worded in different ways is merged into one and counted, so you can see what gets repeated.
The question it answers
“What exactly is AI saying about us?”
From an answer to its claims
Sample data“Fernvale Skincare is a vegan brand. Returns are free within 14 days, and it was founded in 2012.”
- Returns are free within 14 daysSaid in 7 answers
- Founded in 2012Said in 2 answers
↑ opinions aren’t claims
Step two · Look it up
A verdict needs a source.
Each claim is searched on the live web, starting with your own site and then public bodies: government, regulators and official registries. A verdict stands only if its source really came up in that search.
The question it answers
“Is this claim true? Show me the source.”
What each verdict needs
Sample data- Correct
A source found in the search agrees.
- Partly
True, but a condition is missing.
- Wrong
An official source says otherwise. Nothing less counts.
- Couldn’t verify
Nothing official found. Not finding it is never “wrong”.
↑ strict on purpose
Step three · Fix what repeats
Spot what AI gets wrong.
Wrong claims come first, most repeated at the top, each with a link to the official page that says otherwise. Fix it at the source, and the next checks show whether it is still being said.
The question it answers
“What is the worst thing AI gets wrong about us?”
We find the wrong claims and tell you what to fix. We can’t stop an assistant from repeating a mistake.
Wrong claims, most repeated first
Sample data7answers
“Returns within 14 days”
The official page says 30 days.
4answers
“Not sold online”
The official page links the online shop.
2answers
“Founded in 2012”
The official page says 2016.
↑ fix it at the source
Straight answers
We would rather say “couldn’t verify” than guess.
Only official sources can say wrong
A blog or a forum disagreeing with AI isn’t enough. If no official source contradicts a claim, we don’t call it wrong.
Not finding something isn’t proof
Plenty of true things aren’t written on an official page. A claim we can’t confirm is marked “couldn’t verify”, and it stays out of the wrong list.
About you, not your rivals
Verification looks at claims about your brand. It doesn’t make statements about other companies on your behalf.
It looks at answers you already have
It works on the answers in your newest check, so running it uses none of your plan’s answers.
Questions
How is this different from the fact check?
The fact check compares AI’s answers with a fact sheet you write: your own prices, plans and rules. Verification looks claims up on the live web, in official sources. You can use both.
Which plan is it on?
Verification comes with the Medium plan and above. The fact check against your own fact sheet is on every plan.
What counts as an official source?
Your own website first, then public bodies such as government sites, regulators and official registries in the countries that matter for the claim.
Is it always right?
No check is. That is why each verdict links to its source, so you can read it yourself, and why anything without an official source is left out of the wrong list.