Burak Reis

Does llms.txt work? What I learned opening my site to AI

When someone asks an AI for "someone in Istanbul who does brand and motion design," I want my name to come up. Here's what I did about it, and what I learned about llms.txt.

Burak Reis3 min read

The way people find a designer is changing. It used to be that they'd type into Google and look at the sites that came up. Now some of them ask an AI the same question and read a single paragraph of answer. If my name isn't in that paragraph, it doesn't matter how good my site looks. Nobody sees it.

When I built designedbyburak.com, I took that into account. Let me write down what I did, and why one of those things was for nothing.

I opened the door to AI bots in robots.txt

The site's robots.txt file, the file that tells bots which of them can come in, invites AI bots one by one, by name. ChatGPT, Claude, Perplexity, Google's AI bot, all of them.

There's a distinction worth knowing here. OpenAI has two separate bots. One makes you show up in ChatGPT's search; the other crawls sites to train models. Blocking the training bot doesn't affect your visibility in search.1 I left both open. If my writing ending up in a model's training is going to do me some good, there's no sense in shutting that door.

I told the machines who I am, too

In the site's code, in a section visitors don't see but search engines read, it says who I am: my name, what I do, which city I'm in, and my Behance, LinkedIn, Instagram and X accounts. This blog uses the same record. That way a search engine doesn't take the two sites for two different people. It understands that both are the same Burak Reis.

Does this part work? Honestly, it's hard to measure. But it's a method Google describes in its own documentation, and once it's set up it runs by itself.

I learned llms.txt doesn't work

While building the site, I also added an llms.txt file. It's been proposed as a short summary of a site for AIs, and it got talked about a lot over the past year.

Then I looked into it and found this: there's no evidence that the big search engines use the file. Google states clearly that AI features don't need a separate file or any special markup.2 And in June 2025, Google's John Mueller reportedly said that no AI system currently uses it.3

The file is still there; it does no harm. But it's not on this blog. I didn't see the point of adding something I knew wouldn't help.

The real work happens off the site

All of this is just opening the door. What actually gets someone through it is something else. There are studies showing that AI search trusts what other people write about a person or brand more than that person's or brand's own site. I wrote about that in my post on the second-site effect.

So I've made my site readable. Whether my name shows up in those one-paragraph answers depends on my work being talked about somewhere else. That's partly what this blog is for.

Notes

  1. OpenAI, "Overview of OpenAI Crawlers". Source ↩
  2. Google Search Central, guide to AI features. Source ↩
  3. As reported by Search Engine Roundtable, June 2025. Source ↩

Published on November 3, 2026.

Burak Reis designs brand identities and motion. About, portfolio.

Follow via RSS