01It is not one engine, it is several, each with its own crawler
The most common misunderstanding around SEO for AI is treating it as one target. It is not. Google's AI Overviews and AI Mode are built on top of Google's own search index, the same one a normal blue-link result comes from. ChatGPT search works differently: it runs its own crawler, OAI-SearchBot, and also pulls from third-party search providers, Bing among them, so a page can be picked up there without ranking anywhere on Google. Perplexity runs a third crawler again, PerplexityBot. For a small business, the practical consequence is simple: blocking one crawler in robots.txt does not hide you from the others, and optimizing only for Google leaves the other two blind to a page they might otherwise have cited.
- Google's AI Overviews and AI Mode run on Google's own search index.
- ChatGPT search uses its own crawler, OAI-SearchBot, plus outside providers like Bing.
- Perplexity uses its own crawler, PerplexityBot, separate from both.
- A page blocked from one crawler is invisible only to that one engine, not to the others.
02The checklist for a Romanian small business site
Most of this checklist is unglamorous and it is the same work that already helps a site rank normally, applied with a little more discipline. Write the direct answer to the question in the first few sentences of a page, in plain Romanian, before any scene setting. Add real FAQ schema to the pages that already answer questions, so the structured markup matches the visible text word for word instead of contradicting it. Keep the Google Business Profile filled in and current: hours, address, phone and photos, since that is often the first source an AI answer pulls from for a local query. Check robots.txt allows Googlebot, OAI-SearchBot and PerplexityBot rather than a blanket disallow copied from a template. And get the business named, correctly and consistently, name, address and phone the same everywhere, on Romanian directories, local press and partner sites, since that repetition is what lets any of these systems confirm a fact rather than guess at it.
- Answer-first paragraphs in plain Romanian, near the top of the page.
- Real FAQ schema matching the visible questions and answers, word for word.
- A complete, current Google Business Profile: hours, address, phone, photos.
- robots.txt allows Googlebot, OAI-SearchBot and PerplexityBot, not a blanket block.
- Consistent name, address and phone (NAP) across Romanian directories and local press.
03What is not worth paying for
A cottage industry has grown around the fear of missing out on AI search, and most of it sells fear rather than results. Do not pay for a standalone llms.txt package sold as a ranking service, that specific claim is covered in its own post on this blog and the short version is that the file helps a crawler fetch facts, it does not move a ranking. Do not pay for a guarantee that a business will be cited by ChatGPT or Perplexity by a certain date, because nobody controls what a language model chooses to quote, and the mechanics of how a page gets picked are covered in the post on showing up in ChatGPT answers. And be wary of a generic AI SEO audit that reads like it was run on any site, rather than one that names your actual pages, your actual Google Business Profile, and your actual robots.txt.
- Skip a standalone llms.txt package sold as a ranking booster, it is not one.
- Skip any guaranteed citation date for ChatGPT or Perplexity, nobody controls that outcome.
- Skip a generic AI audit that could describe any business, ask for your own pages named specifically.
- Spend the budget instead on the free checklist items: robots.txt, Google Business Profile, FAQ schema.
A rebuild where structured data and local presence did the work
SCP Shine & Style, a beauty salon in Sector 3, Bucharest, had a website already, it just ranked badly and did not reflect the work being done there. PROOF48 rebuilt it with a page per service, VideoObject and Person schema on every clip and blog post, and a landmarks line on every page naming the Dristor and Titan area for local search, on top of redirecting every old WordPress URL so nothing already indexed broke. None of that was built for a language model, it was built for accuracy, and accuracy is exactly what any answer engine, human or AI, ends up rewarding.
See the build: SCP Shine & StyleQuestions, answered
Should I block AI crawlers in robots.txt to protect my content?
Not if you want to be found by them. Blocking OAI-SearchBot or PerplexityBot removes you from exactly the answers you are trying to appear in, while doing nothing to protect content that is already public on the open web.
Is SEO for AI a separate service I need to pay extra for?
Mostly no. Most of the checklist, a correct robots.txt, a filled-in Google Business Profile, FAQ schema, is free configuration and honest content, not a separate paid product, which is why we treat it as part of normal SEO rather than a new line item.
Where PROOF48 fits
SEO for AI is not one target, it is three separate crawlers reading three separate systems, and the checklist that helps with all of them is mostly free: a direct answer up top, real FAQ schema, a current Google Business Profile, a robots.txt that lets the crawlers in, and a consistent name, address and phone across the Romanian web. PROOF48 builds that checklist into every site by default.
Cosmin Tiu · Founder, PROOF48
Cosmin has worked in online marketing and software development since 2018, and has built with AI every day since January 2023. He still writes code from scratch daily, in Python, React, Rust and C#. PROOF48 ships business websites in 48 hours and invoices afterwards, so every opinion on this blog was paid for in real builds, not in theory.