UK High Street Retailers Plagued by AI Chatbot Inaccuracies

The findings highlight a growing disconnect between the information AI models serve to consumers and the reality on the ground. Among the most frequent errors, chatbots misplace retailers into the wrong postcodes in one out of every 10 instances. In particularly egregious cases, the tools suggest addresses located at least 20 miles away from the store's actual site. Beyond location errors, the models often provide dead links to websites or misattribute brand identities.

Searchable co-founder Chris Donnelly notes that smaller businesses are disproportionately affected by these hallucinations. While large corporations benefit from a vast web of third-party data that helps AI models verify their details, local shops often lack a digital footprint extensive enough to correct the systems. Perplexity proved the least reliable in the study, delivering inaccurate information in 10% of cases, while ChatGPT and Gemini recorded error rates of 4% and 5% respectively.

This trend poses a tangible risk to physical storefronts, as consumers increasingly rely on LLMs for product research and store selection. With 58% of shoppers reporting a loss of brand trust when faced with incorrect AI-generated information, retailers are under pressure to bolster their digital presence. Donnelly advises businesses to maintain rigorous updates across third-party directories and leverage local press or customer reviews to provide the training data necessary for more accurate AI representation.