AI Assistants Are Now Choosing Local Businesses for Customers

In the past, when searching for a local business, a user would scan a list of search results, open several options and make the choice themselves. Now, when a user asks an AI assistant for a recommendation, it presents an already curated shortlist. Google calls this AI Mode — the AI-powered search mode — and notes that it is especially useful for complex, comparison-based questions that would otherwise require multiple separate searches.

In its explanation, Google names two core techniques: RAG (Retrieval-Augmented Generation — a method of shaping the model's answer based on pages pulled from an existing search index) and query fan-out — a technique that runs several related searches simultaneously based on a single question. Both are designed to base the model's answer on a broader and more accurate set of sources.

According to an analysis by Whitespark, when testing 540 queries across three US cities and six industries, AI Overviews appeared directly in 15% of queries with local intent, 92% of informational queries, and 97% of mixed queries — such as "do I need to hire a lawyer after a car accident," which combine a purchasing decision within them. It is precisely in this last category of queries that the AI's answer determines which businesses come to the customer's attention.

Uberall's measurements in the quick-service restaurant segment showed that a single query typically yields recommendations for just three to five brands — a figure that illustrates just how narrow this "playing field" is. According to Google's explanation, listings, reviews and reputation all influence whether a business is considered by the AI at all, but meeting these requirements does not guarantee appearance in AI features.

The "Generative AI Performance" report in Google Search Console — not yet rolled out to all sites — shows how many times links to a site's pages appeared in AI features, but it does not show which query triggered the appearance, nor what other assistants such as ChatGPT, Perplexity or Claude say about the business. For this reason, the key recommendation for local businesses is to keep information about business hours, services and address on the website as clear text, consistent across all sources, and to avoid anything that blocks indexing.