In short: AI search tools name brands that are clearly described and often mentioned on the web. What helps: a website with direct answers to typical customer questions, structured data, an llms.txt, open access for AI crawlers and mentions on other sites.
How AI search builds answers
ChatGPT, Perplexity and Google’s AI Overviews search the web and summarize what they find. ChatGPT relies in part on Bing’s search index. If you’re visible on Google and Bing, you already have a head start.
Say clearly what you do
Put one sentence on your site saying what you offer, for whom and where. An AI can pick up “We build Shopify stores for brands in Europe”. It can’t do much with “We rethink brands”.
Answer your customers’ questions
Guide pages with a short answer at the top get quoted often. One question per page, the answer in the first two sentences.
Tech that helps
- Structured data: details about your company, services and articles in the code that machines can read reliably.
- llms.txt: a short text file in your site’s root folder that tells AI services which pages matter. It’s a proposal that not every service reads yet. It does no harm.
- Check robots.txt: if you block crawlers like OAI-SearchBot or PerplexityBot, AI search can’t find you.
- Bing Webmaster Tools: submit your sitemap there too, not just to Google.
Mentions on other sites
AI search trusts brands that others write about: reviews, directories, interviews, named client projects. One good case study on a client’s site counts for more than ten claims of your own.
Frequently asked questions
What is an llms.txt?
A text file at yourdomain.com/llms.txt that describes in plain form who you are and which pages matter. Meant for language models and AI search.
Is this SEO or something new?
Mostly good SEO. What’s new is that clear, quotable answers and mentions on other sites matter even more.
More about websites and landing pages