What Happens When AI Gets Your Brand Wrong

A colleague of mine told me a story a few months ago that I haven't been able to stop thinking about. She runs marketing for a mid-sized professional services firm, and during a routine review of their sales pipeline, she noticed something odd.

A prospect on a demo call said, almost as an aside,

“I asked ChatGPT about you guys and it said you mostly do cybersecurity consulting.”

They don't do cybersecurity consulting. They haven't in years. They sold off that division after an acquisition restructuring, but the old website content was still live on a legacy domain, and apparently enough of the internet still associated them with cybersecurity that the AI had built a confident, entirely outdated picture of what the company does.

The prospect wasn't angry about it, but you could hear the confusion in her voice when she told me. She'd spent six months aligning her messaging, rewriting service pages, and cleaning up their positioning. The machine was still telling people a story about a company that no longer existed.

This is one of the things about AI-driven brand perception that catches people off guard; it's not only about whether you show up or not, but about what the system says about you when you do. And those descriptions are being formed from the full sweep of information available about your company, including the outdated, inaccurate, and contradictory stuff you forgot was out there.

Think about all the surfaces where your company is described online. Your current website, obviously, but also your LinkedIn company page, your Google Business Profile, industry directories you signed up for three years ago and never updated, press releases from before your last rebrand, Glassdoor reviews that reference a division you no longer operate, partner listings that use language from a previous era. Each one of those surfaces is a data point that AI systems use to construct their understanding of who you are.

If all those surfaces tell the same story, the system's understanding is stable and accurate. If they tell different stories, or worse, if some of them tell stories about a version of your company that no longer exists, the system's understanding is going to be messy. Messy descriptions don't just confuse potential customers; they actively undermine the sales conversations your team is already having.

I think most companies underestimate how much digital debris they're carrying around. Old microsites from campaigns that ended two years ago, PDF brochures indexed by Google that describe services you've since discontinued, guest posts on industry blogs written by employees who've moved on that describe the company in terms that no longer apply. None of this stuff is malicious, and most of it was perfectly fine when it was published. But in an environment where AI systems are reading everything and trying to synthesize it into a coherent description, legacy content that contradicts your current positioning is working against you every day.

The fix sounds simple but takes discipline. You have to audit what the internet says about you, not just what you're currently saying about yourself. Those are two very different things, and the gap between them is usually bigger than anyone expects.

This isn't just a large-company problem, either. I've seen small businesses with ten employees discover that their Google Business Profile description says something completely different from their website homepage, because someone set it up years ago and nobody thought to update it when the business evolved. That kind of inconsistency is easy to overlook when your main lead source is referrals, but in a world where an increasing number of people are validating referrals by asking an AI "tell me about this company," even small disconnects create friction.

The thing that makes this particularly tricky is that you don't get a notification when AI systems form an opinion about your brand. There's no crawl report, no index status, no ranking position to check. The system just quietly absorbs information, builds its understanding, and starts describing you to anyone who asks. If that description is wrong, you might not find out until a prospect mentions it offhand during a sales call, the way my colleague's prospect did. By then, who knows how many other conversations happened that you never heard about.

I've started thinking about this as a form of brand maintenance that didn't exist five years ago. We've always had to manage our reputation with customers, with analysts, with the press. Now we have to manage our reputation with machines, and machines don't give you the benefit of the doubt the way a human reviewer might. They don't look at an outdated description and think "oh, they probably restructured since then." They just report what they've absorbed, inconsistencies and all.

The encouraging part is that this is fixable, and in most cases it doesn't require a massive budget or a specialized agency. It requires the tedious, unsexy work of finding every place your company is described online and making sure the description matches reality. It requires retiring content that no longer applies rather than letting it sit there accumulating influence in the wrong direction. It requires treating your digital footprint as something that needs active maintenance rather than something you set up once and forget about.

None of that is glamorous work, but the companies I've watched get this right share a common trait: they treat their online presence like a living thing that needs regular attention, not a monument they built once and walked away from.

I go much deeper into the mechanics of how AI systems form these brand perceptions, and how to systematically fix them, in my book Explainable: Why AI Recommends Some Brands & Ignores Others.