Yes, AI search deserves a small, deliberate share of a local business owner's attention. ChatGPT Search is available across ChatGPT plans and can use location for nearby recommendations. Google has also built AI answers into Search. The route exists. It usually shouldn't displace work that already brings calls, bookings or visits.

A restaurant, specialist retailer or design-build firm may be compared inside an AI answer before the customer opens a map. A towing company with the wrong phone number on Google has a more immediate problem. Protect the customer paths that already work, then test AI search with a short set of real buying questions.

Start with the decision your customer is making

AI search matters most when a customer needs help choosing. OpenAI's own research on consumer ChatGPT use found that practical guidance and information-seeking are among its largest uses. The data says ChatGPT often helps people think through a decision. It cannot tell you whether people in your town use it to find your trade.

Look at the purchase instead. A customer comparing kitchen remodelers may ask about style, budget, permits and older homes in one conversation. Someone with a burst pipe at 1:30 a.m. is more likely to need a working phone number and a nearby plumber who answers. The first decision gives an AI assistant room to compare. The second punishes delay.

Referrals do not remove the issue. A referred customer may still ask ChatGPT whether the company looks legitimate, serves their area or has experience with the job. In that case, accurate facts and proof matter more than being named in an unprompted list of providers.

  • More attention is justified when customers compare several credible options before contacting anyone
  • Location, specialty, timing or customer fit changes the right answer
  • A light check is enough when repeat customers and trusted referrals already drive most new work
  • Wrong hours, weak service pages or missed calls are costing business now and should be fixed first

The evidence is enough to test

ChatGPT Search can rewrite a question into targeted web searches, use general or precise location and show sources beside an answer. Google says AI Mode is built for questions involving further research and comparison. Both products can place a business in front of someone who has not yet chosen whom to call.

The local evidence is still narrow. An August 2026 preprint checked 2,208 search-grounded answers against 4,776 restaurants, cafes and bars in Canggu and Ubud, Bali. Across ChatGPT, Claude, Gemini and Perplexity, 85.6% of the venues were never recommended. The systems also disagreed sharply about which venues belonged in their top 20.

That study found associations between entering an answer and having an owned website, more reviews, listed prices and more third-party web mentions. It did not prove that any one change caused visibility. Food and drink in two Bali markets cannot stand in for a Canadian roofer, dental practice or accounting firm. For an owner, the paper supports a limited conclusion: local recommendations happen, many legitimate businesses are absent and the platforms disagree.

Choose now, monitor or wait

Make a proportionate decision for this quarter instead of predicting the future of search. Use the state that best matches the business today, then write down what would cause you to change it. My rule is simple: AI search earns more time only after a repeated check finds a customer-facing problem.

Run one short check before setting a budget

Write four questions a real customer could ask. Use the actual service, town and constraint. ‘Who is the best contractor?’ is too loose to guide a decision. ‘Which contractors around Moncton handle century-home kitchen renovations while the owners remain in the house?’ gives the answer something concrete to judge.

Run the same questions in two fresh ChatGPT Search conversations and once in Google AI Mode. Save the date, location wording, businesses named, reasons given and visible sources. You are looking for a repeated customer-facing problem, not a flattering screenshot.

  • One discovery question that does not name your company
  • One comparison question using a real customer constraint
  • One question that asks whether your company fits that job and why
  • One factual check for services, service area, hours and contact details

Spend first on work that survives a platform change

If the check exposes a problem, favour work that also helps customers and ordinary search. Correct the hours on the website and the profiles people use. Explain the service in plain language. Publish the evidence a careful buyer would want before making contact. Earn honest reviews through the normal customer follow-up.

Google says the same SEO fundamentals apply to its AI features. A page must be indexed and eligible for a search snippet before it can appear as a supporting link, and there is no special AI file or schema required. OpenAI likewise says allowing OAI-SearchBot makes a site available for inclusion, but does not guarantee placement.

Be wary of work that has no value outside a promised AI result: dozens of thin articles, new directory profiles nobody uses, schema fields the visible page cannot support or daily prompt checking. The platform may change before those tactics pay back. A clear service page, current business record and credible proof still have a job tomorrow.

Set a review date and return to the business

If the first check finds a material accuracy problem, fix it and retest after the corrected pages have been recrawled. If the answers are accurate but mixed, save the evidence and check the same questions next quarter. A daily watch will mostly show normal answer variation and consume attention better spent on customers.

Write the next check date beside the four questions. Until then, keep asking new customers how they found you. A named ChatGPT referral is stronger evidence for your business than a general argument about where search is heading.

Sources

  1. OpenAI: ChatGPT Search
  2. OpenAI research: How People Use ChatGPT
  3. Google Search Central: AI features and your website
  4. Invisible to the Machine: Auditing AI local recommendations