Features · Brand perception
Brand perception: what AI says about you, and whether it's true
GeoRush brand perception analyzes how AI assistants talk about your brand when they mention you: the tone on each engine, the themes you are praised or criticized for compared with your competitors, and the specific claims they repeat (pricing, dates, features), checked against your official facts to catch errors.
Updated on
The problem it solves
Being mentioned by AI is only half the story. The other half is what it says. An assistant can recommend you and, in the same sentence, claim you “don't offer a free plan” when you do, or keep associating you with “slow support”. Buyers read that as fact and quietly cross you off.
GeoRush reads every mention and shows you the tone, the themes and the specific facts AI repeats. When a claim contradicts your official facts, it is flagged as false and GeoRush helps you correct it at the sources AI draws from.
How it works
- 1
Add your official facts
What is true about your brand, one fact per line, with numbers: “From $12 per user per month”, “Founded in 2018”, “Integrates with Shopify”. The more precise they are, the better the checking.
- 2
Claude analyzes every mention
It classifies the tone (positive, neutral or negative), extracts the themes linked to your brand and your competitors, and collects sentences with specific facts about you.
- 3
Claims are checked against your facts
Each claim is marked as correct, false, or unverified (when there is no official fact to compare it with).
- 4
False claims become tasks
False claims and recurring criticism go into the Action plan, with steps to fix them and a ready-to-paste prompt for your AI.
What you see in GeoRush
- Net sentiment: positive minus negative mentions over the total. Above +30% is a solid reputation.
- Tone per engine: how each assistant talks about you when it names you.
- Themes: you vs your competitors, with tone; for example, AI linking you with “price” and a rival with “ease of use”.
- Praised for / criticized for, with a recommendation for each case.
- What AI claims about you, with a verdict (correct, false or unverified) and how often each claim is repeated.
- Mentions per engine: how often each assistant names you and in what tone.
How to read the fact-check
| Verdict | What it means | What to do |
|---|---|---|
| Correct | The claim matches your official facts. | Nothing urgent. Keep that fact visible and current. |
| False | The claim contradicts one of your official facts. | Publish the correct fact on your site, update directory and review profiles, request corrections where it's repeated, and add it to llms.txt and JSON-LD. |
| Unverified | There is no official fact to compare it with. | If it matters, add the relevant official fact so it can be checked. |
Where errors come from
False claims about a brand usually come from outdated pages, directories with old information or articles nobody has revisited. So the fix isn't “talking to the AI”; it's fixing the sources. Start with your own site, with the fact clearly visible and in structured data, then the third-party sites AI cites, which you can find in Sources and citations.
Your official facts are reused elsewhere in GeoRush too: the outreach email Claude drafts only describes your brand with them, and the llms.txt draft includes them.
What does AI say about your brand?
Create a free account, add your official facts and check whether AI describes you correctly.
Frequently asked questions
Keep reading
Competitor analysis
Who AI recommends instead of you, engine by engine, why each rival wins, and newly discovered brands.
Sources and citations
Which sites AI cites in your market and how to get listed on them, with contact finder, pitch and pipeline.
Action plan
Tasks ranked by estimated impact, with steps, copy-paste code and before/after measurement.
What is llms.txt
What llms.txt is, how to write one with a ready example, and what to expect from it, without the hype.