Why is this happening more often now?
Customers increasingly ask ChatGPT, Gemini or Perplexity before they ask you: what a service costs, whether you offer something specific, whether you are still open. Most of the time the answer is accurate and the customer arrives pre-sold. Sometimes it is not, and a customer walks in or calls quoting a price from six months ago, a service you discontinued, or a location you closed, stated as plain fact. A clinic or salon that updates its price list can keep fielding the old number from customers for months afterward, if the old figure still lives on a directory page nobody updated. Your sales-floor staff need a way to handle that moment, because it is becoming a normal part of the conversation.
What should staff say in the moment?
- Do not argue with "the AI". The customer is not wrong to trust it, and defensiveness reads as dodging.
- State your current fact plainly and specifically: "our price for that service is X as of this month" beats "that's not right".
- Offer the honest reason if you know it: "we updated that six months ago and some directories are still showing the old rate".
- Give them the corrected information in writing, a WhatsApp message or email, so they are not relying on memory of a verbal correction.
- Log what happened. One customer mentioning it is a signal; three in a month is a pattern worth investigating.
How do you turn the correction into trust instead of friction?
Most customers don't mind that a mistake exists somewhere online. What they're gauging is whether you'll handle it honestly when it surfaces. A calm, specific correction, backed by something concrete like an updated price sheet or a quick screen-share of your current listing, usually closes the gap faster than the AI error opened it. Treat the moment as a chance to demonstrate you are on top of your own information.
What do you do after the customer leaves?
Trace the error to its source rather than treating it as a one-off. Ask which tool the customer used and what exactly it said, then check that answer yourself across all four major engines. If the wrong fact keeps surfacing, follow the correction path for wrong AI information: find the stale source feeding it, fix it there, and confirm the fix later rather than assuming it stuck. A pattern of these moments is one of the clearest signals that your AI visibility needs a proper audit rather than one-off corrections.