Damian Mansell runs The Plastics Shed, a building-plastics supplier in the UK. Earlier this month he discovered that Google's AI Overview had summarised his business as overwhelmingly disliked by customers. The trouble was, most of the negative sentiment it had scraped together belonged to rivals and to unrelated firms that sell literal garden sheds made of plastic. Mansell had been paying Google roughly £700 a month in ads while its own AI quietly told searchers his business was bad news. It took weeks of submitted corrections to get it fixed, and Google refused to refund a penny of the ad spend.
He is not an isolated case. An Idaho property manager found her company merged with a similarly named firm that had closed years ago. An SEO consultancy was branded a scam over an old Reddit thread about a completely different app. As AI Overviews and chatbot answers become the first thing a customer sees before they ever reach your website, a growing number of small businesses are discovering that the AI got it wrong, and that there is no obvious button to press to fix it.
This matters because the stakes are no longer hypothetical. Trustmary's monitoring of AI visibility across dozens of businesses found that of those tracked for more than three weeks, 65% managed to shift how often and how favourably they appeared in generic AI searches, changes that moved rankings from the bottom of a list to "first" or "most popular" within weeks. AI answers are volatile and responsive to the data feeding them. Unfortunately, that volatility runs in both directions: a business can climb quickly, or it can be quietly buried by someone else's bad press.
1. Audit what AI is actually saying about you
Before you can fix a problem, you need to know it exists. Open ChatGPT, Google AI Overviews, Perplexity and Gemini, and ask the kind of question a customer would ask: "Is [your business] any good?", "Best [your trade] in [your town]?", "What do people say about [your business name]?" Note who else gets mentioned in the same breath. If a rival, a similarly named company, or an old forum thread keeps surfacing next to your name, you have found your problem before it costs you a customer.
2. Understand why the mix-up happens
AI Overviews are built by scraping and summarising whatever text mentions a name that resembles yours, with limited ability to check whether "The Plastics Shed" and a plastic garden shed retailer are actually the same business. Common name patterns, franchise-style branches, and businesses that have rebranded or closed are especially prone to being folded together. The AI is not maliciously targeting anyone. It is doing statistical pattern matching without accountability for getting it wrong, and currently, small businesses carry all the risk of that trade-off.
3. Correct it, and document everything
Google, and other platforms, do allow business owners to flag inaccurate AI summaries, but the process is slow and inconsistent. Screenshot the incorrect summary with a timestamp. Submit a correction through Google Business Profile support and any relevant feedback tool in the AI product itself. Keep a dated record of when you flagged it and when, if ever, it changed. This paper trail matters if you later query ad spend charged while the false information was live, even though refunds are not guaranteed.
4. Build a first-party signal AI cannot easily confuse
The most durable fix is not chasing corrections after the fact, it is giving AI models a clean, well-structured, unambiguous source of truth about your business before confusion sets in. That means keeping your reviews centralised, current, and clearly linked to your exact business name and location, with structured data such as review schema markup that helps AI attribute reviews correctly. A steady stream of recent, specific, first-party reviews acts as a stronger and more current signal than scattered mentions elsewhere, and it is far harder for an AI system to misattribute a review that is clearly structured and dated against your verified profile.
| Step | What to check | How often |
|---|---|---|
| Audit | Ask ChatGPT, Gemini, Perplexity and Google AI what they say about your business | Monthly |
| Compare | Note any rival, defunct, or unrelated business mentioned alongside yours | Monthly |
| Correct | Flag inaccurate summaries via Google Business Profile and in-product feedback tools | As soon as found |
| Document | Screenshot and date every incorrect AI summary and correction request | Ongoing |
| Reinforce | Keep reviews recent, centralised and structured with your correct business details | Weekly |
A quiet cost most owners never see
The most unsettling part of Mansell's story is not that the AI got it wrong. It is that he was paying to advertise against his own reputation without knowing it. Most small business owners have no routine for checking what AI says about them, which means the mistake can run for months before anyone notices the drop in enquiries. Given that AI-generated answers are only going to become a bigger part of how customers first encounter a business, treating a monthly AI audit with the same seriousness as checking your Google reviews is no longer optional. The businesses that get ahead of this will not be the ones with the best product. They will be the ones who noticed first.