How Orange County service businesses get recommended by ChatGPT and AI answers
More people are starting the hiring search inside an AI now. Instead of scrolling ten blue links, a homeowner in Irvine types "who’s a good HVAC company near me" into ChatGPT, or asks Google’s AI Overview for the best plumber in Costa Mesa, and reads the short answer it gives back. If your business isn’t in that answer, you didn’t lose the customer to a competitor with a better ad. You lost them before the ad ever loaded.
The good news is that getting cited by an AI isn’t a mystery or a trick. These systems pull from a handful of signals you can actually control, and most local service businesses fail on the same three. Here’s what they are, in plain terms, and how to check your own site.
1. The AI has to be allowed to read your site
When ChatGPT or Perplexity answers a "who should I hire" question, it often fetches live web pages in that moment to ground its answer, then cites the ones it used. Those fetches come from named crawlers. If your site blocks them, you can’t be cited, full stop.
Here’s the part almost nobody knows: there are two different kinds of AI crawler, and they’re not the same decision.
- Search and citation crawlers fetch a page when someone asks a question, so the AI can quote it and link to it. These are the ones you want to let in: OAI-SearchBot and ChatGPT-User (ChatGPT), PerplexityBot, ClaudeBot, and Google-Extended (Google’s AI answers).
- Training crawlers like GPTBot and CCBot scrape pages to help train future models. Letting these in is optional and it’s a values call, not a visibility one. You can block training and still be fully citable in the live answers.
Most sites that block AI don’t do it on purpose. A developer pastes a restrictive robots.txt from an old project, or a security plugin flips it on, and the search crawlers get shut out along with everything else. The check: open yourdomain.com/robots.txt in a browser and look for Disallow lines under the crawler names above. If you can’t tell, that’s the answer, and it’s worth having someone look.
2. The page has to answer the question, not dance around it
AI engines quote passages that stand on their own. A paragraph that says "we’re passionate about serving our community with quality solutions" gives the AI nothing to lift. A paragraph that says "we do same-day AC repair in Costa Mesa, Newport Beach, and Irvine, and we answer the phone 24/7" is a clean, quotable answer to a real question.
Write the way a customer asks. Put the direct answer first, in one self-contained sentence or two, before the story and the sales language. If a page can’t be quoted in isolation, it won’t be.
3. The AI has to be sure you’re a real business
An AI is cautious about recommending a name it can’t confirm. It’s effectively asking: is this a real, findable business, and do the details line up? You confirm it by being consistent everywhere it looks. Your name, address, and phone number should match exactly across your site, your Google Business Profile, and any directories. Structured data (a bit of code that labels your business, services, and area for machines) helps. So does an llms.txt file, a plain-text summary of your site written for AI systems to read.
None of this works, though, if the pages themselves aren’t even in Google’s index. A page Google never crawled can’t be found by the AI answers that lean on Google either. That’s the floor, and it’s more common to be sitting below it than owners expect.
On a roughly 4-month-old domain, 8 of 8 revenue pages that were "unknown to Google" got indexed, and the site is set up so the AI search crawlers can read and cite them.
Run the check on your own site
- Open
yourdomain.com/robots.txtand confirm the AI search crawlers aren’t blocked. - Read your top service page out loud. Is there a sentence in the first paragraph that directly answers what you do and where? If not, that’s the fix.
- Search your business name in Google. Does one clear, consistent listing come back, with a phone number that matches your site?
- Paste a key page URL into Google Search Console and confirm it’s actually indexed, not "unknown to Google."
If any of those turn up a problem, that’s not bad news. It’s visibility sitting on the table that costs almost nothing to claim, and it’s the exact work we do for Orange County service businesses. If you’d rather someone just handle it, that’s what WEO is for.