How to Get ChatGPT, Gemini and Perplexity to Find and Recommend Your Business

More and more customers ask an AI assistant “who can build me a website in Mexico?” instead of opening Google. This discipline is called GEO (generative engine optimization) or AEO. The good news: much of it overlaps with good SEO. The bad news: nobody can guarantee that a model will mention you, and anyone who promises it is selling smoke.

How AI assistants find information

There are two paths. The first is the knowledge the model was trained on, which is updated every so often and is hard to influence. The second is live search: when the assistant consults the web to answer, which is why your site can be read and cited. That's where you can do real work.

• Google (AI Overviews and AI Mode) relies on its own index. Google has said no special markup is needed: SEO fundamentals apply.
• ChatGPT with search uses its own crawler (OAI-SearchBot) and, according to OpenAI, external search providers.
• Perplexity crawls the web with PerplexityBot and shows its sources.
• Claude can also search the web and respects the robots.txt rules for its crawler.

1. Let them read you

Check your robots.txt file. If you block AI crawlers by default, you won't appear in their answers. These are the most relevant ones:

• OAI-SearchBot and ChatGPT-User (OpenAI): search and browsing requested by users. GPTBot is the one OpenAI uses for training.
• PerplexityBot (Perplexity).
• ClaudeBot, Claude-SearchBot and Claude-User (Anthropic).
• Google-Extended: controls whether Google uses your content for its Gemini models; it doesn't affect your appearance in Google Search.
• Applebot-Extended (Apple).

You decide which ones to allow. For a business that wants customers, it usually makes sense to allow the search and citation crawlers. Also check that your CDN or firewall isn't blocking them by mistake.

2. Deliver the content in the HTML

Many AI crawlers don't run JavaScript the way a browser does. If your content only appears after scripts load, your page looks empty to them. Server-rendered sites (like the ones we build with Next.js) deliver the full text from the very first moment.

3. Write to be cited

• Answer the question in the first lines, with a clear, direct sentence.
• Use headings that are real questions from your customers and answer them underneath.
• Include concrete facts: what you do, for whom, where, in how much time, what's included.
• Avoid filler language; models extract sentences that carry meaning.
• Stay consistent: your name, description, phone and location should be the same on your site and on your profiles.

4. Structured data

Mark up your organization (Organization or ProfessionalService), your articles (Article), your frequently asked questions (FAQPage) and your local business details if applicable, using JSON-LD. Google has restricted FAQ rich results to certain sites, but the markup still helps machines understand your content.

5. Build mentions outside your site

Models tend to recommend entities they see mentioned in several trustworthy places. Look after your Google Business profile, LinkedIn, industry directories, verifiable reviews, interviews, guest articles and, if you meet the criteria, Wikidata. Reputation is built over time, not with a technical tweak.

6. Show up on Bing too

Several assistants query Bing indexes. Register your site in Bing Webmaster Tools, submit your sitemap and consider IndexNow to announce changes instantly.

What about the llms.txt file?

It's a proposal to offer models a summary of your site in a text file. As of this article, we know of no public confirmation that the major providers use it to decide what to cite. It costs little to create, but it shouldn't be your priority or your strategy.

How to measure whether it works

1. Ask several assistants about your service and your city, worded in different ways, and note whether they mention you. Repeat every month.
2. In your analytics, look for visits referred from chatgpt.com, perplexity.ai, gemini.google.com or copilot.microsoft.com.
3. Check your server logs to see which AI crawlers visit your pages.