The Orchard ChatGPT Sent Customers to on the Wrong Day

Dana Kessler grows apples on forty acres in Hadley, Massachusetts, land her grandfather cleared in 1962. Every fall her farm stand runs u-pick weekends through late October. Last October, a family of five drove in from Springfield after asking ChatGPT a simple question: "family apple picking near Northampton this weekend." ChatGPT named Kessler Family Orchard. The family showed up on a Saturday morning. The gate was chained. Three days of rain had turned the rows to mud, and Dana had posted a closure on her Google Business Profile at seven that morning. ChatGPT had not seen it.

AI search engines like ChatGPT, Perplexity, and Google AI Overviews name a business by combining several sources, then picking whichever one looks most current and most consistent with the rest. They are not reading your website in real time the way a person clicks a link. They cache summaries, some updated daily, some updated every few weeks. If your Google Business Profile says "closed for rain" but a two-year-old blog post about your orchard still ranks well and says "open daily through October," the engine may quote the older, more established page instead of your live status. This is not a bug aimed at small businesses specifically. It is how these tools balance speed against freshness. The fix is not a new website. It is making sure the same hours, closures, and facts appear everywhere your business is mentioned, updated the same day something changes, so there is no older version left to compete with the truth.

Why did ChatGPT give my customer the wrong hours?

ChatGPT did not invent the wrong hours. It pulled them from an older source, probably a blog post or directory listing, that still ranked well and had not been updated. When two sources disagree, generative engines tend to trust whichever one looks more established, not whichever one is most recent.

Dana traced her problem to a regional tourism blog that had written up her orchard back in 2019. The post still said "open daily, rain or shine." Her own site said almost the same thing, because she had copied her original hours from that same post years earlier. Two matching, wrong sources beat one correct, fresh one.

She had also never claimed a listing on two smaller apple-picking directories that show up when people search "apple orchards Massachusetts." Those blank profiles did not hurt her directly. But they meant fewer places online agreed with her current hours, and agreement is part of how these engines decide what to trust.

Where do AI search engines actually get local business facts?

They pull from a mix of your Google Business Profile, Yelp, Facebook, local directories, review sites, and any page that mentions your business by name. No single AI engine visits your live website every time someone asks. They rely on cached, indexed versions, refreshed on their own schedule, not yours.

That is why a business can be named correctly and still be wrong in the details. Being found is not the same as being accurate. In BrightLocal's Local Consumer Review Survey, 98 percent of consumers said they had used the internet to find information about a local business within the past year, often checking more than one listing before they trusted it (BrightLocal, "Local Consumer Review Survey" (2023), https://www.brightlocal.com/research/local-consumer-review-survey/). AI engines run a faster version of that same cross-checking, automatically, before they ever answer a customer's question.

Google's own Business Profile guidance tells owners to keep hours current and notes that Google may show information pulled from other public sources when a business's own profile is not the clearest answer available (Google, "Google Business Profile Help" (accessed 2026), https://support.google.com/business/). That single line explains what happened to Dana's family from Springfield.

What Dana changed

She started with the tourism blog. A phone call and an email got the old post flagged as outdated, with a note linking to her current hours. She claimed the two blank directory listings and filled them in word for word to match her Google Business Profile: same phone number, same hours, same closure language.

Then she built a habit. Every time the weather closes the orchard, she updates her Google Business Profile first, within minutes, before she posts to Facebook or texts her part-time staff. That order matters. It is the source AI engines tend to check first for a physical location like hers.

She has not had another family show up to a locked gate since. That does not mean every AI answer about her orchard is now perfect. It means the sources feeding those answers finally agree with each other, and with reality.

The short version

Getting named by ChatGPT or Google AI Overviews is only half the job. The other half is making sure every listing that mentions your business says the same true thing, updated the same day it changes. Stale sources do not disappear on their own. They just keep answering for you until you catch up to them.

If you want a clear picture of what AI engines are currently saying about your business, and where the mismatches are, Ozvor checks that for you at ozvor.com.