45% of consumers globally said they asked an AI tool like ChatGPT for a local business recommendation in the past year, up from just 6% the year before. If your restaurant's online presence is a logo, a stale menu PDF and one glossy shoot from last season, there is nothing for ChatGPT to recommend.
42% — More than 42% of consumers globally were already using AI tools to assist with shopping tasks like product discovery and comparison as of July 2026, according to industry data cited by Square.
2026 — As of July 2026, U.S. restaurants taking orders via ChatGPT or Claude through Square's new integration pay only Square's standard 2.9% plus $0.30 per-transaction processing fee, with zero marketplace commission, compared to DoorDash's 15%-30% and Uber Eats' 20%-30% commission tiers.
7 — For AI-generated orders that require delivery, Square's white-label dispatch network charges U.S. restaurants a flat courier fee of roughly $7 to $10 per order, rather than a percentage-based commission, as of July 2026.
45% — Globally, 45% of consumers said they asked an AI tool such as ChatGPT or Claude for a local business recommendation in the past year (2026), up sharply from just 6% the year before.
49% — In the United States, 49% of adults said they use AI chatbots like ChatGPT, Gemini or Copilot in 2026, up from 33% in 2024, according to Pew Research Center.
24% — In the United States, about 24% of adults reported using AI chatbots on a daily basis as of February 2026, more than double the roughly one-in-ten who used them at all in prior years.
AI is already a discovery channel, not a novelty
This is not a fringe habit anymore. In the United States, 49% of adults said they use AI chatbots like ChatGPT, Gemini or Copilot in 2026, up from 33% in 2024, according to Pew Research Center. About 24% of adults reported using AI chatbots daily as of February 2026, more than double the roughly one-in-ten who used them at all in prior years.
This kind of habit does not stay confined to one country. More than 42% of consumers globally were already using AI tools to assist with shopping tasks like product discovery and comparison as of July 2026, according to industry data cited by Square. Restaurant recommendations sit inside that same behaviour. People type a question into an assistant instead of scrolling Google, and the assistant answers with whatever it can find and trust.
What ChatGPT is actually recommending you on
An AI assistant does not have a private opinion of your restaurant. It synthesises what is already published across your Google Business Profile, Instagram, TripAdvisor, your website and food blogs. Consistency across those places is what tells the model your details are real.
A restaurant with the same name, the same menu items and matching photos of actual dishes across five channels gets picked up with confidence. A restaurant with mismatched hours, an outdated menu on the website and a feed that has not moved in months looks unreliable, even if the food is excellent.
Here is the trap. One annual photoshoot gives you a burst of matching content for a few weeks, then nothing new for the rest of the year. The model has no fresh signal to work with in between. Consistent, well-tagged photo and menu content across every channel, updated often, is what actually feeds these systems.