What is ChatGPT saying about your brand?

Key takeaways

• Engines describe you to buyers before those buyers ever reach your website. That description is the first impression now.
• The description is assembled from third-party sources more than from your own site, because engines weight what they did not get from you.
• Competitor comparison pages are a common source. Your competitor gets to characterise you, and the engine repeats it.
• Wrong pricing is the single most frequent error, and the most expensive one.
• Publishing a rebuttal does not work. You have to change what the engine is reading.

Why this is different from traditional reputation management

Classic online reputation management is about what ranks when someone searches your name: reviews, news, a bad forum thread on page one. The buyer reads the sources and forms their own view.

AI reputation management is about what an engine says, in its own words, having already read those sources and formed a view on the buyer’s behalf. There is no page one to manage. There is a paragraph, and it is generated fresh each time.

That changes the work. You are no longer pushing a bad result down; you are correcting the inputs to a summary. And because engines prefer sources they did not get from you, your own website is often the least influential thing you control.

Where wrong descriptions come from

• Competitor comparison pages. A rival publishes "us vs you", characterises you unflatteringly, and the engine reads it as a source about you. This is the most common and most annoying case: your competitors choosing what gets said about you.
• Stale pricing. An old plan page, a retailer listing, or a review site that captured your price two years ago.
• Abandoned positioning. An old press release or directory listing describing a product you no longer sell.
• Thin or aged review profiles. A handful of old reviews carries more weight than it should when there is nothing newer.
• Directory and aggregator entries nobody has updated since they were created.

How to audit what engines say about you

1. Ask each engine directly: what is [brand], who is it for, what does it cost, how does it compare to [competitor]. Use ChatGPT, Claude, Perplexity and Google AI Mode.
2. Record the exact wording, not a summary of it. The specific phrasing tells you which source it came from.
3. Ask the engine for its sources where it will give them, then read those pages.
4. Check the facts hardest: price, positioning, target customer, what you do and do not do.
5. Repeat monthly. Descriptions drift as sources change.

How to actually change it

In rough order of effectiveness:

1. Correct the source. If a directory, retailer or review profile carries the wrong price, update it. This fixes the answer at the root and is usually the fastest win available.
2. Get into better sources. Editorial roundups and credible third-party coverage outweigh your own site. If the only thing written about you is written by a competitor, that is what gets quoted.
3. Publish the facts in a machine-readable form. Clear, current, self-contained statements of price, positioning and fit, with schema, so an engine does not have to infer them.
4. Build review recency. Volume matters less than freshness on the platforms engines actually read.
5. Answer the comparison question yourself. If a competitor owns the "you vs them" page, an honest version from you gives the engine an alternative source.

What this is worth

A buyer asking an engine about your category reads three or four sentences about you and decides whether to keep looking. If those sentences carry a price you abandoned a year ago, or a competitor’s characterisation of your weakness, the shortlist forms without you and you never see it happen. There is no bounce rate for an answer you were left out of.

This is also the part of AI search almost nobody sells. Most agencies work on presence. Very few have a practice for accuracy, which means it is usually the cheapest unclaimed advantage in a category.

Where to start

Run the audit above yourself. It takes half an hour and you will usually find at least one wrong fact worth fixing that afternoon.

If what you find is systemic, that is what Ante does, at a published $6,000 a month, alongside the work of being named in the first place. The free AEO report covers the same ground if you would rather see the picture before talking to anyone.

Frequently asked questions

What is AI reputation management?

The work of controlling what AI engines say about your brand, rather than only whether they mention it. Engines assemble a description from sources across the web and present it as an answer, so an out-of-date price or a competitor’s characterisation can become the first thing a buyer reads about you.

How do I find out what ChatGPT says about my company?

Ask it, in the way a buyer would. What is [brand], who is it for, what does it cost, how does it compare to [competitor]. Do the same in Claude, Perplexity and Google AI Mode, record the exact wording rather than a summary, and ask for sources where the engine will give them. The phrasing usually reveals which page it came from.

Why is ChatGPT saying the wrong price for my product?

Almost always because a source it trusts carries the old number: a retailer listing, a directory entry, a review site or an archived plan page. Engines prefer third-party corroboration, so your own updated pricing page may lose to a stale listing elsewhere. Find and fix the source rather than publishing a correction.

Can I make an AI engine remove something about my brand?

Not directly, and be sceptical of anyone claiming otherwise. What you can do is change what it reads: correct the underlying sources, add credible current ones, and make the accurate version easy to find and easy to parse. Answers update as sources update, though not instantly.

My competitor wrote a comparison page about us. Does that affect AI answers?

Yes, frequently. Engines treat a competitor comparison as a source about you, and if it is the only substantial page comparing you, its framing tends to survive into the answer. The practical counter is to publish an honest version yourself and to get independent third parties writing about the same comparison, so the engine has something else to weigh.

How is this different from traditional online reputation management?

Traditional ORM manages what ranks when someone searches your name, and the buyer reads the sources themselves. AI reputation management concerns what an engine says in its own words, having already read those sources for the buyer. There is no results page to reorder, only a generated paragraph whose inputs you can change.

How often should I check what AI engines say about my brand?

Monthly is a reasonable cadence for most companies. Descriptions drift as sources are updated, republished or outranked, and a single reading tells you very little. What matters is the trend and whether corrections you have made have actually propagated.