A customer can now ask an AI assistant to find a business like yours, check your price and your reviews, and book the job, without ever opening your website. Your site still decides that answer, because that is where the assistant reads your facts. Its new job is to state them clearly enough for a machine to act on.
Think of what a customer asks you on the phone. Where do you work, are you licensed, what do you charge, how do customers rate you, can you come tomorrow. An assistant has to answer the same questions, and it pulls them from your pages and your schema markup, the hidden code that spells out your details for machines, plus your Google Business Profile, your Yelp listing, and reviews on other sites. Your website is the one piece you fully own, so everything else should match it.
This is where pages written to rank fall down. “Reliable electrical services” gives an assistant nothing to match. A page listing your services, your service area, your hours, your licensing, and recent reviews gives it enough. A restaurant may get asked for family owned, vegetarian and vegan options, wheelchair access, an accessible restroom, and a table for six tonight. A general description cannot answer that.
Outdated prices are the trap. They can linger on your own site, in your schema markup, and in third party discussions like Reddit threads. AI engines increasingly treat your website as the hub for cross checking what they find elsewhere, so a stale number carries weight.
Ten minutes a day
- Open ChatGPT or Gemini and ask what your customers ask before they choose a business. Leave your name out of the prompt, because a branded question gives you only a narrow view of what customers see before they have made a choice.
- Check whether you appear, whether your details are accurate, which competitors show up, and which sources the assistant relies on.
- Note whether it recommends you or only mentions you. Those are not the same thing.
Keep measuring clicks and keep working your existing funnel. Some customers now choose you based on what an assistant told them, and those decisions may never show up in your analytics.
One caveat. Machine attention is not customers. A 90 day Bluehost analysis counted 83.4 million verified AI crawler requests a day across the sites it hosts, roughly 90% of them for model training, about 9% for search indexing, and roughly 1% for live assistant fetching. For ChatGPT, the report estimated one observable referral visit for every 5,350 verified HTML crawl requests. That is not a conversion rate, and being crawled is not being chosen.
