How AI Assistants Know Your Business Hours and Prices in 2026 (and How to Fix Wrong Info)

Read this with AI

Read with ChatGPT Read with Claude Ask Gemini

Short answer: In 2026, AI assistants like ChatGPT search, Google AI Overviews, Perplexity, and Microsoft Copilot read your hours and prices from your website, Google Business Profile, Bing Places, Yelp, and directory listings, then repeat whatever source they trust most. When one of those is stale, the assistant repeats the wrong facts. Fix the sources in priority order.

A nail and beauty salon owner in Milpitas, California, watched this happen in real time. On the OpenStreetMap community forum, they wrote that "the business hours keep changing and it show up incorrectly no matter how I arrange the hours." Their map pin was "still tagged from a previous Sprint/T-mobile store," so on Apple devices "my business location doesn't even show up on their phone or Instagram map." In their own words: "It has been extremely difficult and it affects the business drastically because they can't find the business."

That is the whole problem in one story. A non-technical owner, real lost customers, wrong hours and a dead former tenant's data being repeated across maps and AI answers, and no idea which source to fix first. This article is the map. It covers where the facts come from and how to correct them. If you want the separate question of why an assistant recommends one shop over another, that is the ranking logic, covered in how ChatGPT decides which businesses to recommend. Here we only care about one thing: making the facts it states about you true.

Where do AI assistants get your business hours and prices?

They do not know your hours. They repeat them. Every assistant is reading from a short list of public sources and stitching an answer together:

Your own website (especially your contact page and any structured data on it)
Your Google Business Profile
Your Bing Places listing
Yelp
Other directories and aggregators (Apple Maps, data providers, industry listings)
The model's own training data, which is a frozen snapshot from months or years ago

The order matters, and it is different for each assistant. Google's AI features lean on the Google ecosystem. ChatGPT leans on Bing. When two of these sources disagree, the assistant picks one, and it does not always pick the fresh one. That is why a bakery can update its holiday hours on Google, still get quoted the old hours by ChatGPT, and never understand why.

Why does ChatGPT lean on Bing and Bing Places?

Because it is built that way. OpenAI's own help documentation for ChatGPT search says the feature partners with third-party search providers, and it names Bing, from Microsoft, explicitly. When you ask ChatGPT about a local business, it rewrites your question into targeted search queries, sends those to providers like Bing, and shares your general location, based on your IP address, to localize the results.

Read that again, because it changes your to-do list. If ChatGPT is asking Bing, then your Bing listing is feeding ChatGPT. Most owners have never once looked at Bing Places. They optimized Google and stopped. The same Bing dependency is why Microsoft Copilot answers the way it does, since Copilot sits on the same search stack. If Copilot is a channel you care about, the Copilot-specific playbook goes deeper on it.

What feeds Google's AI, Yelp, and the directories?

Google's AI Overviews and AI Mode draw local facts from the surfaces Google already controls: Maps and Search. And the hours that show on Maps and Search are the hours you set on your Google Business Profile. Google's own support documentation is blunt about this: you edit your business hours on your Profile, and those are the hours that display for your business across Google. So your Profile is not one lever among many for Google's AI. For hours, address, and phone, it is the lever. If you want the fuller version of Profile as an AI-visibility tool, see using your Google Business Profile for AI search.

Yelp and the big directories are their own separate feeds. Perplexity and other assistants will cite Yelp, TripAdvisor, or a trade directory directly when those pages rank well for your name. Each one is a source the assistant can read, and each one can be wrong on its own. A price you changed two years ago can still be sitting on a directory page nobody told you existed.

Why does AI keep repeating your old hours and prices?

Two reasons, and they stack.

The first is the training snapshot. Every model was trained on a frozen copy of the web from some past date. If the assistant answers from memory instead of doing a live search, you get facts from whenever that snapshot was taken. That is why an assistant can confidently give you a price you have not charged since last year.

The second reason is worse, because it survives even live search: stale upstream data that outranks your fresh data. This is exactly what happened to the Milpitas salon. Their map data was still tagged as a former Sprint and T-Mobile store, and that stale record kept winning over the owner's own corrections. Bing describes the same trap in its own documentation. Bing populates a business listing from many internal and external sources, and while Bing Places is one of the primary sources, other sources can take priority if their data looks fresher, for example when your Bing Places data has not been updated for more than six months. Let your listing sit untouched and a third-party aggregator can overwrite you.

In our work correcting these listings for owner-run shops, the pattern is almost always the same. The owner fixed Google months ago and assumed the job was done. The wrong number the assistant keeps quoting is coming from a source they never claimed: an old Bing record, a Yelp page with a former address, a directory that bought its data from an aggregator in 2023. The fix is never "argue with the AI." The fix is finding the specific stale source and correcting it there.

Do not underestimate how often the machine is simply wrong. A Columbia Journalism Review Tow Center study tested eight AI search engines across 1,600 queries and found they returned incorrect information in more than 60 percent of cases. Assume the assistant can be wrong about you, then go check.

How does structured data label your hours and prices for machines?

Structured data is a small block of code on your own site that states your facts in a format machines read without guessing. It uses the Schema.org vocabulary, and the relevant type for a shop or service is LocalBusiness.

Two properties do the heavy lifting. The `openingHours` property lets you write your weekly hours in an unambiguous format, using two-letter day codes like Mo and Tu and 24-hour times, so "Mo-Fr 09:00-17:00" means exactly that with no room for a machine to misread it. The `priceRange` property lets you state your price level, for example `$$`, so an assistant reporting "how expensive is this place" has a source to quote instead of a guess. This is the one source you fully control, so it is where correction starts. For the wider picture of how schema shapes what AI and Google display about you, see how structured data powers knowledge panels.

Start with your own website and its structured data

Your site is the only source you own outright, so make it correct and machine-readable first. Put your real hours, address, phone, and price range in plain text on your contact page. Then add LocalBusiness structured data with `openingHours` and `priceRange` filled in to match. When an assistant crawls your site, it now has a clean, unambiguous version of the truth straight from the source, and a consistent record it can cross-check the messier third-party listings against.

Fix your Google Business Profile hours and price range

Sign in to your Google Business Profile and set your regular hours, plus any special hours for holidays. Google's documentation confirms these are the exact hours that show across Maps and Search, which are the surfaces Google's AI Overviews and AI Mode read from. This single edit corrects the largest slice of what Google-based AI says about you. While you are in there, confirm your address, phone, and category are current too.

Claim and update your Bing Places listing

This is the step almost everyone skips, and it is the one that fixes ChatGPT and Copilot. Go to Bing Places for Business, claim your listing, and update your hours, address, and phone. Because Bing feeds ChatGPT search, a correct Bing listing is how you get ChatGPT to stop repeating old facts. And because Bing deprioritizes listings left untouched for more than six months, this is not one-and-done. Log back in a couple of times a year so your record stays the fresh one. If Bing itself is a channel worth more attention, showing up on Bing covers it in full.

Correct Yelp and the big directories

Claim your Yelp page and fix the hours and price range there, then work through the other directories that show up when you search your own business name. Apple Maps, TripAdvisor, and industry-specific directories each hold a separate copy of your facts, and any one of them can be the source an assistant quotes. If your map pin is wrong, as it was for the Milpitas salon, this is also where you correct the underlying location data, including the aggregators that feed Apple Maps.

Make your name, address, phone, and hours match everywhere

This is the step that keeps the wrong facts from coming back. NAP consistency means your name, address, phone, and hours are identical across every source: your site, Google, Bing, Yelp, and the directories. When they all agree, there is no stale version left to win, and no conflict for an assistant to resolve in the wrong direction. Suite written two ways, an old phone number on one listing, last year's hours on another: each mismatch is a door for stale data to walk back in. Consistent listings are also a local ranking signal, covered in local citations for small business.

Frequently asked questions

How do I make ChatGPT update my business hours?

Fix your Bing Places listing, then your website. ChatGPT search runs its live queries through Bing, so a corrected Bing record is the most direct path. You cannot edit ChatGPT directly, and you cannot argue it into changing its mind. You change the source it reads, then wait for it to re-crawl.

Why does AI show a completely different business at my address?

Almost always because a former tenant's data is still attached to your location in an upstream source, exactly like the Milpitas salon whose pin was still tagged as an old Sprint and T-Mobile store. Find the stale source, usually a map data provider or directory, and correct the record there rather than only updating Google.

How long until the correct information shows up?

It varies by source. Your own site updates as soon as an assistant re-crawls it. Google and Bing edits can take days to propagate to the surfaces AI reads. Directories are slowest. Make every correction, then re-check in a couple of weeks, because assistants are frequently wrong and worth auditing more than once.

Should I check what AI already says about me first?

Yes. Ask ChatGPT, Google, Perplexity, and Copilot for your hours, address, and prices, and write down every wrong answer before you fix anything. That tells you which sources are stale. For a repeatable way to run that audit, see how to check what ChatGPT says about your business.

The Milpitas owner's mistake was not carelessness. It was not knowing that the wrong hours lived somewhere they had never looked. Assistants do not invent your facts, they repeat them, so the whole job is making sure every source they can reach agrees with the truth. Fix the sources in the order above, keep them consistent, and the answers follow. That constant syncing across your site, Google, Bing, Yelp, and the directories is exactly the busywork Fonzy handles for owner-run shops, so the assistants quote you correctly without you chasing each listing by hand.

Sources

Roald

Founder Fonzy. Obsessed with scaling organic traffic. Writing about the intersection of SEO, AI, and product growth.