The new Google is destroying small businesses' reputations

Damian Mansell was stunned to see what Google was saying about his nascent business. When he searched for reviews of The Plastics Shed — the online building-plastics supplier he incorporated at the start of 2025 — the platform's AI overview said customer feedback was "overwhelmingly negative." It listed complaints about delayed deliveries, lying staff, and damaged products. While the company had some positive feedback, the summary said, its poor customer service was a "significant recurring issue." The good news: The reviews weren't actually about Mansell's company. They appeared to be for competitors and companies that sold actual plastic sheds. The bad news: He had no idea what to do about it.

Mansell, who lives in the UK, says. "I can see how that happens, but why should it happen with Google?" Mansell, who lives in the UK. "I mean, obviously, the forefront of AI technology."

To make matters worse, Mansell was paying Google about £700 a month to advertise, while the summaries warned people away. Following advice he found on online forums, he repeatedly submitted feedback to Google that the abstract was wrong. After a couple of weeks, it started to improve. AI has been an invaluable tool for him to build out his business, so he doesn't want to malign the tech in general, but he wishes there were more accountability when things go awry.

"AI gives the common man the knowledge, but it can also ruin the common man," he says.

Google's AI Overviews are rapidly becoming consumers' first impression of businesses. Instead of scanning reviews and websites, many users see a single synopsis that purports to blend information from across the internet into a comprehensive digest. When those summaries are inaccurate, misleading, or jumbled, business owners say they can cause serious reputational damage and financial losses, and there's often little recourse.

Mansell still wonders how much business the issue cost him. He's tried to get Google to refund some of his ad dollars, but he hasn't had any luck, despite his best efforts.

"I'm like a dog with a bone," he says. "But I met my match with Google."

Internet search has changed drastically in the last few years: Instead of a list of links, Google often provides a single response that's supposed to summarize the constellation of online information. These overviews look right and sound confident, but they draw on a litany of sources with varying levels of reliability. They can even spit out information that is flat-out wrong. (Like recommending people make glue pizza.) Misinformation on the internet isn't new, but the clean, concise, AI-concocted package is.

"With the traditional search engine, the user sees multiple things. So if one of them has something wrong, chances are something else would counterbalance it," says Chirag Shah, a professor at the University of Washington's Information School. That's now gone, and with AI overviews, you're getting "The Answer," which may or may not be correct.

I look at the AI overview, I wouldn't call me.

For business owners, the consequences can be more than a mild annoyance or temporary confusion. These AI summaries are becoming their digital storefronts. Across the internet, you can find entrepreneurs and managers grumbling that AI summaries mix them up with other companies, surface complaints that are directed at someone else, or dole out false facts.

Earlier this year, Betty Whitney started noticing that Google's overviews were conflating her company — NW Select Property Management in Idaho — with similarly named businesses. The top panel seemed to be merging her firm with one in the region that closed years ago and mixing her reviews up with property managers in other states.

"If I'm a customer, and I'm looking for a property management company, and I look at the AI overview, I wouldn't call me," she says.

Whitney has spent months trying to amend the situation. Like Mansell, she's used Google's feedback mechanism to give overviews a "thumbs down" when they're wrong, and she's made some adjustments to her website to try to feed the AI crawlers more accurate data. After bringing her problem to a Google support forum, she got in touch with a third-party SEO expert who was able to help her out — sort of. The overview has gotten better, but it still periodically reverts to the mistake-filled version.

Three years after relocating, Philippa Main, a real estate agent in Northern Virginia, still can't completely convince Google that she's no longer in Florida. When she searches her name, most of the information that comes up about her is correct, but then there's a line that confidently states that she's been "servicing the Tampa Bay area since 2014," even though it sits right above her Virginia address. Main's tried everything she could think of to get it adjusted, combing the internet to try to find where the AI is drawing from, emailing Google, and asking friends to report the issue.

"There's so much competition in my industry that if any single thing seems off, someone's just going to call the next person on the list," she says. It's especially frustrating for small businesses, because "we're just trying to do everything that we can to compete with these massive companies who actually do have direct lines to Google or their representatives," she says. "Google just doesn't seem to care."

In a statement, a Google spokesperson told me that its search-related AI experiences are "rooted in our quality ranking systems and are designed to present a range of perspectives" from all over the internet. "AI Overviews are responsive to people's specific queries; for example, if someone specifically searches for complaints about a business, the generated response will likely show relevant information from sources across the web," they said.

It's no secret that AI is not always a bastion of truth — almost everyone who's used the technology has experienced a response from it that's highly off-base at some point. As the New York Times wrote in April, Google processes over five trillion searches a year, and even if its overviews are right nine times out of 10, that still means half a trillion wrongs. Google acknowledges that while the overwhelming majority of its overviews are accurate, there can be cases where they miss context or misinterpret content. A Google spokesperson said the study the Times cited has "serious holes."

Google's summaries synthesize information from many places — a big source is, obviously, the company website, but it also gobbles up Reddit posts, 10-year-old blogs, and Yelp reviews. The platform provides links that are supposed to back up its claims, but those links don't always support the output. AI has also been known to hallucinate, meaning it invents plausible-sounding things from thin air.

"There's so much room for error," says Lily Ray, an SEO and AI search consultant and the founder of Algorythmic, a consultancy.

The overviews are delivered with such assurance that people don't realize they're looking at an extracted or generated answer that may be incomplete or incorrect. Instead of clicking on five links to compare information or just spending a few minutes confirming, they skim the automated summary and call it a day. The AI says this roofing company's reviews are terrible? Onto the next one! Rarely do people dig in to check if it's pulling complaints for a business in another state.

There's so much room for error.

Search industry professionals say this is a new frontier for businesses. They no longer have to focus so heavily on search rankings but must instead manage the AI's interpretation of their reputations. It's not about chasing clicks —it's about making sure AI knows you exist and is nice and correct about you. Ray says it's the "biggest change to search" she's seen in her 16-year career.

This brave new world presents all sorts of nuances and complications. Google likes to cite Reddit a lot, which "can go awry very fast," Ray says. Reddit has a lot of good information, but it can also be a little wild. The same goes for YouTube comments, which the AI also seems to like. Some brands suffer from an information void: there's not a lot of content out there about them, so AI tries to fill in the gaps or comes up with bad answers. Or, they've got a name problem where they're too close to another entity, and the model can't tell who's who. There may be bad actors who intentionally leave false or negative information about businesses online for AI summaries to pick up. Even simple facts, such as store hours or phone numbers, require a concerted effort across the entire internet to keep straight. "There's so much maintenance work that has to go into keeping a brand's content and information accurate and up to date," Ray says.

Michael King, the founder and CEO of iPullRank, a digital marketing agency, tells me he focuses on citation accuracy and on gaining some influence over AI outputs. "The way these systems work is they're basically doing a bunch of searches in the background, and then they're feeding content to the large language model," King says. Businesses need to create more "surface area" — meaning publishing more content and targeting more keywords — to help AI find the right answer. He encourages clients to position themselves as the experts on their own brands.

"You've got to think of it as more like a reputation management campaign than your classic SEO campaign," King says. "It's just far more multidimensional."

Ben Fisher does this for a living and still runs into problems. He noticed that Google's AI summary was warning that his company — Steady Demand, a SEO and social media consultancy for small businesses — was a scam. After doing some digging, he realized it was referring to an old Reddit thread about a similarly named app and had to take some time out to "train" Google to know the difference.

"The big problem is there's nobody to contact. The other big problem is you search once, and you're done," Fisher says. Large language models don't produce the same results every time, even for the same questions, so people don't realize one result might not match the next month, week, or minute. "It's still a situation where you should be monitoring things on a regular basis," he says. "Otherwise, you're just not going to know why you're not getting calls."

This is a difficult issue to tackle from a technological, entrepreneurial, and regulatory point of view. LLMs are improving, but they're never going to be perfect. Business owners can do their best to keep an eye on how they're showing up in search results, but they've also got 9,000 other things to do.

Reasonable minds — and different countries and legal systems — can disagree about how responsible Google should be when the robot screws up. In Canada, a musician has filed a $1.5 million lawsuit against Google claiming that its AI summary falsely identified him as a sex offender. A court in Germany recently made a preliminary ruling that Google is liable for false statements made in its AI overviews. A Google spokesperson said that the German case focuses on "specific and narrow errors," not the way overviews display content, and that the company disagrees with the ruling and plans to appeal. Shah says that in the US, we have "very little consumer protection" for these types of issues.

"I don't think lawmakers even fully understand the technology enough and the implications to be able to do anything," he says.

In the meantime, business owners are left white-knuckling it, hoping that the mysterious technology at the heart of those AI summaries looks kindly upon them. That's the case with Mansell, who's proud to say that Google's overview of the Plastics Shed is now "fantastic", just like many of his actual reviews.

"It does worry me with regard to what can be said about you without any recourse," he says.

Despite his frustrations with the summary and failed attempts to get a refund, Mansell still pays to advertise with Google — otherwise, people don't click through to his website. "I just gave up," he says. "It was just an absolute pointless exercise."

Emily Stewart is a senior correspondent at Business Insider, writing about business and the economy.