Something new is happening in how customers find businesses. Instead of typing a few keywords into Google and scrolling through ten blue links, more and more people simply ask — "Who's a good life insurance advisor in Los Angeles?" "What's the best accounting software for a small law firm?" — and get a short, confident answer from ChatGPT, Perplexity, Gemini, or Google's AI overviews.
Here's the part most business owners haven't caught up to yet: those tools recommend specific companies. By name. With a phone number. We've watched it happen for clients — someone asks an AI a question and is told to call a particular business. So the natural question is: how do you become the business the AI recommends? The good news is it isn't magic, and it isn't pay-to-play. It comes down to a handful of things you can actually check and fix. Here's the roadmap.
1. Make sure the AI crawlers can even reach your site
AI tools learn about your business by reading the web, using automated crawlers with names like GPTBot and OAI-SearchBot (ChatGPT), PerplexityBot, ClaudeBot, Google-Extended, and CCBot. If your site quietly blocks these bots, you're invisible to the tools they feed — no matter how good your content is.
The first place to look is a small file at yoursite.com/robots.txt, which tells bots what they're allowed to read. Plenty of sites block AI crawlers here without the owner ever realizing — often a plugin or security service did it automatically. Confirm the AI bots aren't disallowed, and while you're at it, check that your firewall or CDN (like Cloudflare) isn't turning them away at the network level, which robots.txt won't reveal. It's a five-minute check that decides whether anything else even matters.
2. Put your best content on your own website — not just someone else's
This one trips up the most businesses. A lot of owners have written genuinely good, helpful articles — but published them on Medium or LinkedIn and just linked to them from their own site.
The problem: when an AI reads that article, the credit largely goes to Medium, not to you. The content isn't on your domain and isn't tied to your business as the entity the AI would recommend. You did the work; someone else's website earns the authority. If that describes you, the highest-value move is to bring that content home — republish it as real pages on your own site. Done properly, with the right technical tags, this creates no "duplicate content" problem; you're simply telling search engines which version is the original.
3. Make sure your content is actually readable — not trapped in PDFs or images
AI tools read text. They struggle with anything locked inside a PDF, a scanned image, or a graphic. If your most useful material — a case study, a guide, a fee schedule — only exists as a downloadable PDF, it's effectively invisible to the models. Anything you want an AI to be able to quote should live as real, selectable text on a normal web page. It's a simple format change that can move a key piece of content from "can't be read" to "gets recommended."
4. Structure your pages the way people actually ask questions
People talk to AI in full questions: "How much life insurance does a business owner need?" "What does a fractional CFO actually do?" The tools reward content structured to match — clear question-style headings with direct, useful answers underneath.
Most business websites are written as marketing prose ("Our team delivers tailored solutions…"), which reads fine to a human but gives an AI little to match a real question against. You don't have to rewrite everything — often it's enough to reshape what you have under plain-English question headings and add an FAQ section that answers the real questions customers ask. An FAQ is practically purpose-built for how these tools work.
5. Add the behind-the-scenes labels (schema markup)
Schema markup is invisible code that spells out, in a language machines understand, exactly what your page is: this is a business, here's its name and location, this is a service it offers, this is a question and here's the answer. It's like putting clean labels on everything instead of making the AI guess. Most sites have none, or only the bare minimum their theme added by default. Adding proper Organization, Service, and FAQ markup helps AI tools understand who you are and what you answer — and confidently pass that along.
6. Build trust signals off your site, too
AI tools don't just read your website — they weigh how you show up across the wider web. A complete, accurate Google Business Profile, consistent name, address, and phone details everywhere you're listed, genuine reviews on the platforms these tools read, and mentions in reputable third-party sources all tell the AI you're a real, established, trustworthy business worth recommending. Testimonials on your own site are nice; corroboration elsewhere carries more weight.
7. Measure it, so you know it's working
You can't improve what you don't watch. Start by simply asking the AI tools the questions your ideal customer would ask, and note whether you come up — and who does instead. Do it across ChatGPT, Perplexity, Gemini, and Google's AI overviews, since each pulls from different sources. Re-check periodically. And when a new lead comes in, ask how they found you; "ChatGPT told me to call" is a data point worth tracking.
The short version
If you want AI tools to send you leads, here's the checklist:
- Confirm AI crawlers aren't blocked in your
robots.txt(or by your firewall/CDN). - Get your best content onto your own website, not just Medium or LinkedIn.
- Move important content out of PDFs and images into real page text.
- Restructure pages around real questions, and add an FAQ.
- Add proper schema markup so machines understand your pages.
- Strengthen your off-site presence: Google Business Profile, reviews, citations.
- Test the AI tools regularly and track how new leads find you.
None of these are exotic. They're the same fundamentals that make a website work well for humans — clear, readable, well-organized, and trustworthy — applied to a new kind of visitor that happens to be a machine reading on a customer's behalf. The businesses that sort this out now will be the ones getting recommended while their competitors are still wondering where the leads came from.