Preparing Your Site for AI Search
AI models have to be able to read a brand before they can mention it. On most sites, the problem isn’t the content — it’s access and format: the crawler is blocked, the page structure doesn’t tell a machine what it’s looking at, or the text simply isn’t written in a citable form.
This service builds that technical layer. Writing content is a separate job — what I do here is make your existing content visible and citable to AI.
What I set up
Crawler access. I check whether AI crawlers like GPTBot, ClaudeBot, PerplexityBot, and Google-Extended can actually reach your site. On most sites they’re blocked without anyone realizing it — sometimes a security plugin, sometimes a server setting. Which ones to allow and which to block is your call; I explain the options and their consequences.
An llms.txt file. A new standard that tells AI, in a single file, what your site is about and which pages matter most. I generate it automatically, so it keeps itself current as the site grows.
Entity schema. Structured data that defines you and your business in a connected, consistent way: who you are, what you do, what you’re an expert in, which platforms you’re on. It’s the machine-readable answer to AI’s question of “what is this brand.”
Citable content structure. Models don’t quote long, winding paragraphs — they quote clear question-and-answer blocks, definition sentences, and structured lists. I review and edit your existing pages with that lens.
Q&A markup. Making the FAQ sections on your pages also readable as question-and-answer on the machine side. It looks small, but it directly affects the odds of being cited.
The proof is this site itself
I applied every point I’ve described here to this site first. Right now, on the site you’re reading:
oguzerol.com/llms.txthas an auto-generated file listing 300 pages- Person, business, and site are all defined in a single connected schema graph
- 285 pages carry structured question-and-answer markup
- AI crawler access is open and audited
I care about setting something up on myself before I sell it. If you want to see it, I can walk you through the site’s source code and schema output line by line.
How long does it take, and what happens after?
Technical setup is usually one to two blocks of work, depending on the size of the site and its infrastructure. WordPress, Astro, custom software — it doesn’t matter, it applies to all of them.
Setup alone doesn’t put you into AI answers, it makes you eligible to appear in them. I say that plainly because plenty of people promise otherwise. If you want to see where you currently stand, check AI Visibility Audit; if you want to track the change after setup, see Brand Tracking in AI Search.
How do we start? I take a quick look at your site and tell you which items are missing — this first look is free. After that I send the scope and timeline in writing. You can reach me from the contact page.
Frequently Asked Questions
Doesn’t allowing AI crawlers mean letting my content get stolen?
It’s a legitimate concern, and the decision is yours. There’s a distinction worth making: crawling for model training and live access for generating an answer are different things, done by different crawlers. It’s possible to opt out of training while staying open to appearing in answers. I explain which crawler does which, and you decide.
Does Google use llms.txt?
There’s no official statement that Google uses it — that’s the honest answer. Right now it has more traction with other AI tools. Since it costs little to set up, I think it’s worth doing, but I don’t sell it as “required for Google.”
I can add schema markup with a plugin — what’s the difference?
Plugins usually stamp disconnected schema onto every page. What matters isn’t isolated markup, it’s a connected graph: the person tied to the business, the business tied to the site, the page tied to the author. Disconnected schema doesn’t give AI a consistent definition of the entity. Most of what I do here is building those connections.
My site isn’t WordPress — can you still do this?
Yes. WordPress, Astro, Next.js, custom software — it applies to all of them. This site itself moved from WordPress to Astro, and the same structure was rebuilt there too, so I’ve done both firsthand.