The way we search for information is undergoing a profound transformation. As major language models like ChatGPT, Gemini, and Perplexity weave themselves into the fabric of our daily routines, we are increasingly turning to them for quick answers, including vetting local businesses. However, a sobering study by the firm Searchable has revealed a significant disparity in how these AI tools represent businesses, casting a shadow of doubt over their reliability. After conducting over 13,000 tests across 165 London-based firms, the researchers discovered that a staggering 93% of these businesses were described inaccurately. This isn’t just a minor glitch; it is a fundamental breakdown in the technology’s ability to act as a trustworthy digital directory.
When you peel back the layers of these findings, the performance gap between corporate giants and smaller, independent businesses becomes impossible to ignore. Small and medium-sized enterprises (SMEs) are bearing the brunt of these inaccuracies, with half of them receiving at least one piece of false information in the AI’s response. In comparison, larger companies held a much steadier ground, facing a lower error rate of 32%. Even more concerning is the tendency for AI to “hallucinate”—essentially inventing facts out of thin air—where SMEs are more than twice as likely to have non-existent information fabricated about them. When a user relies on an AI to find a service, a business’s very identity is effectively being rewritten by an algorithm that doesn’t know better.
The confusion extends beyond simple factual errors; it strikes at the core of brand identity. The study found that AI tools are prone to misattributing names and confusing smaller brands at a much higher frequency than they do for major, household-name corporations. This places smaller firms in a precarious position. When a potential customer asks an AI for a recommendation or contact details, they expect precision. Instead, they are often met with “discovery gaps”—missing phone numbers, incorrect founding dates, or a failure to list the essential services that keep a small business afloat. When these foundational details are fumbled, it isn’t just an inconvenience; it’s a barrier that keeps local customers from finding the businesses they need.
The root of this problem lies in how these sophisticated AI models are built. They are trained on a massive sprawl of public web data, a digital landscape that is naturally biased toward entities with the loudest voices. Major brands benefit from years of press coverage, thousands of backlinks, and established directory listings, making them highly visible landmarks in the AI’s training data. Conversely, a local shop or a boutique service provider often has a much smaller digital footprint. As Chris Donnelly, co-founder of Searchable, aptly puts it, if an AI doesn’t see a business frequently mentioned in its training set, it either makes up a fill-in-the-blank answer or hits a dead end. For the small business owner, this looks suspiciously like lost revenue walking right past their front door.
This situation highlights a concerning reproduction of historical digital inequalities. Large firms have long dominated traditional search engine rankings through sheer marketing muscle, and there is a very real fear that AI could simply reinforce this “incumbent-first” reality. When an AI constantly prioritizes or correctly identifies the big players while fumbling the facts for the independent players, it risks freezing the market in place. If prospective customers are told that a smaller, equally capable business doesn’t offer a specific service or doesn’t have a valid phone number, they will almost always gravitate toward the competitor that the AI identifies with confidence, regardless of which firm is truly the better choice.
Despite these challenges, the situation is far from hopeless. The beauty of this new AI-driven era is that the rules of engagement are still being written. Unlike traditional search engines, which are often locked into rigid patterns of authority, AI models prioritize structured, accessible, and high-quality data. By optimizing their digital presence to be more “AI-readable,” smaller businesses have a genuine opportunity to level the playing field. The data suggests that if SMEs improve how they represent their services and identity across public records and platforms, they can bridge the visibility gap. The goal is to move from the current state—where they are invisible or misidentified—to becoming reliable, verified, and essential entries in the next generation of discovery.