Your Customers Are Asking AI About You Right Now
Picture a restaurant owner who gets a message from a regular. She asked ChatGPT for late-night dinner spots nearby, and it told her his place closes at 10 pm. It actually closes at 1 am. Nobody on his team had any idea the AI was saying that.
Most owners have not noticed this yet. People now ask ChatGPT, Gemini, Perplexity and Google’s AI Overviews where to eat and which store to buy from. Those answers are not your website. They are a summary of whatever pages the model trusts.
On November 19, 2026, Duane Forrester argued in Search Engine Journal that the next big shift at Google may not be an index update anyone announces. He points to a Google Research paper from October 27 that describes a “sleep” style process, where a model folds new information into its long-term memory on a cycle. If systems like that ship, what AI believes about your business could change quietly, on a schedule nobody tells you about.
The practical lesson is simple. Check what AI says about you on a routine, the same way you check your Google reviews.
Why This Hits Small Businesses Hardest
AI answers are wrong more often than people assume. When the Tow Center for Digital Journalism at Columbia tested eight AI search tools across 1,600 queries, more than 60% of the responses were inaccurate. That test used news articles, not brands, but the lesson carries over.
A big brand has a PR team watching this. A restaurant, mall operator or Shopify store usually has nobody. The wrong detail is rarely dramatic. It is old hours, a dropped delivery area, a changed return policy or a discontinued product.
Each one quietly sends a buyer elsewhere, and you never see that lost sale in your analytics.
What the Top Guides Cover and What They Skip
I read the articles currently ranking for this topic. Most give you a solid testing method. Build a fixed list of 20 to 40 buyer questions, run them in fresh sessions with memory turned off, repeat each one a few times and log who gets named.
That part is useful. But most of those guides come from vendors selling a tracking tool, and they assume you have a content team. Almost none explain how a restaurant or online store actually gets a wrong answer corrected. That is where we will spend our time.
The 30 Minute Monthly AI Answer Audit
This is the routine we use with clients. Once the sheet is set up, one person can run it in about half an hour.
Step 1. Write 10 questions your customers would ask
Skip the keyword tools. Open your Google reviews, Instagram DMs, WhatsApp chats and support emails, and pull out the real questions people ask before buying.
Mix three types. Some should name you, like “Is [your store] legit?” Some should name your category and city, like “best biryani delivery in Clifton.” Some should be comparisons, like “[your brand] vs [competitor] for kids clothing.”
Step 2. Run them in a clean session
Use a private window or log out first, so past chats do not shape the answer. Ask each question in ChatGPT, Gemini, Perplexity and a normal Google search. Run your most important questions twice, because answers shift.
Step 3. Log every claim in a simple sheet
Make five columns for the question, the platform, whether you were named, what it said about you and which links it cited. Then mark each fact as correct, wrong or outdated.
The cited links matter most because they show which pages the AI trusted. The wrong fact usually traces back to an old directory listing, a stale Facebook page or an old blog post.
Step 4. Fix the source, not the chatbot
You cannot email ChatGPT and ask for a correction. You fix the pages it reads. Work in this order.
- Your own website first. Publish one plain, dated facts page with your hours, locations, delivery zones, policies, and what you sell. Update the date whenever something changes.
- Google Business Profile next. Check hours, holiday hours, categories and your menu or product links. For restaurants and malls, this one profile feeds a lot of what Google’s AI shows.
- Every other listing after that. Foodpanda, Tripadvisor, Facebook, your Instagram bio, marketplace profiles and local directories. Contradictions between these are where AI gets confused, so make them match as closely as you can.
- Product pages for online stores. On Shopify or WooCommerce, mark discontinued products clearly or redirect them, keep shipping and return pages current, and make sure your product schema shows real price and stock status.
Step 5. Make sure AI crawlers can reach you
Check your robots.txt file and any firewall or CDN settings. If OAI-SearchBot, PerplexityBot or Googlebot are blocked, those tools cannot read your corrected pages. They keep repeating old information from other sites instead. We often find accidental blocks from overcautious security plugins.
Step 6. Re-check in two to four weeks
Run the same questions again, word for word, so results stay comparable. If a wrong fact is still showing, check the cited links again. Something upstream has not been updated yet.
A Scenario We See With Shopify Stores
Here is a pattern we run into a lot. Picture a kids clothing brand on Shopify whose new customer sales have slowed while ad spend and rankings look steady. When the owner runs this audit, Gemini and Perplexity both tell shoppers the store only ships within one city.
The store has shipped nationwide for over a year. The cited sources turn out to be an old press feature and an outdated Instagram highlight. The fix takes one afternoon. Update the shipping policy page with a clear, dated line, rewrite the About page, refresh the Instagram highlight and ask the publication to update one sentence.
Timing is never guaranteed, because each AI tool refreshes on its own schedule. But once the sources agree, later re-checks usually show the right answer. No new ads and no redesign. Just cleaner facts.
What Not to Waste Money On
You do not need an expensive AI visibility tool to start. A spreadsheet and 30 minutes a month will catch the wrong facts that actually cost you sales. Paid tracking makes sense later, once you run dozens of questions across several markets. Tracking your answers over time also tells you far more than guessing at update dates.
Frequently Asked Questions
How often should a small business check what AI says about it?
Once a month is enough for most businesses. Check again within a week of any big change, like new hours, a new location, a price change or a new shipping policy.
Can I ask ChatGPT or Google to correct a wrong answer directly?
Not in any reliable way. Feedback buttons exist, but the lasting fix is updating the pages the AI cites, starting with your website and Google Business Profile, then your other listings.
Do I need special schema or code for AI search?
Not to start. Consistent facts on your own pages matter most. For online stores, accurate product schema with price and availability is worth adding, and most Shopify and WooCommerce themes support it with small changes.
The Takeaway
AI assistants are already describing your business to customers you never meet. Spend 30 minutes a month checking what they say, fix the sources they cite and keep your facts consistent everywhere. It is one of the cheapest ways to protect sales right now.
If you want a hand with this, whether that means building or fixing a Shopify or WooCommerce store, developing a custom app, or getting unstuck with ads and SEO, our team at ConceptRecall is always up for an honest conversation.