What AI SEO actually means
AI SEO is the work of making a business visible inside AI-generated answers: Google’s AI Overviews and AI Mode, ChatGPT, Gemini, Perplexity and Copilot. It is not a separate discipline with separate rules. Google’s own guidance says its generative features run on the same core ranking systems as regular search, and that pages must be indexed and eligible for a snippet before they can appear in AI features at all.
Google published that guidance in May 2026 and updated it in July; you can read it on Google Search Central. Most of the tactics sold as “AEO” or “GEO” either restate that page or ignore it.
How an AI answer gets built
Assistants use retrieval-augmented generation. The question is broken into several related searches, a set of candidate pages is fetched, and the answer is assembled from what those pages say, with citations. That creates two separate gates: retrieval (can the system fetch and understand your page) and selection (is your page the clearest source for the specific claim). Failing the first gate is silent. There is no error; you are simply not in the candidate pool.
The retrieval gate is technical: crawler access, real HTML rather than JavaScript-only rendering, and topical authority. The selection gate is editorial: does the page say the thing directly, with specifics, and does it agree with what your other pages and profiles say. Our generative engine optimization page has a crawler check you can run on your own site.
What the data says about AI traffic
Seer Interactive’s study of 53 brands, 5.47 million queries and 2.43 billion impressions found that when an AI Overview appears, organic click-through drops sharply, but pages cited inside the Overview keep far more of their clicks than pages that are not cited. Semrush’s analysis found roughly 62% of AI citations come from pages outside the top ten organic results, which is the single most encouraging number for a small business: you do not need to be number one to be quoted.
Why local businesses are better positioned than they think
Nearly everything written about AI search is aimed at national brands competing against publishers with enormous authority. A local plumber is competing against a few dozen other plumbers, and the signals that decide a local answer are the ones a small operator can control: business information that agrees with itself everywhere, real reviews, and content about the actual service area rather than the metro name repeated.
Our founder’s own company is named by Gemini and Grok on national accessible-bathing queries, with larger manufacturers absent. The case study shows the answers as they appeared.
The seven changes that move the needle
- Question-shaped H2s with a direct answer first. Two or three sentences that answer the question, then depth. Assistants quote the sentence, not the section.
- Facts stated as facts. Prices, service areas, hours, years in business, credentials. Stated identically on the site, the Google Business Profile and the schema.
- Entity consistency. One exact business name everywhere. Our post on entity signals covers it.
- Structured data that matches the visible page. Organization or LocalBusiness with a real address, plus FAQ and Service markup where it is true.
- Crawler access. Check robots.txt for the retrieval bots: OAI-SearchBot, ChatGPT-User, Claude-SearchBot, PerplexityBot. Blocking them removes you from citations.
- Real reviews with specifics. A review that names the service and the town is data an assistant can use.
- Being mentioned elsewhere. Brand mentions correlate more strongly with AI visibility than backlinks do. Directories, the chamber, local press, the BBB.
What does not help
Google says so directly: special schema written “for AI,” Markdown-only versions of pages, chunking content into AI-shaped pieces, and seeding fake mentions do nothing for its generative features. Hidden text written for machines is treated as hidden text, which is spam. A file called llms.txt is not used by Google Search; it may help other systems, and it should never replace crawlable pages and a sitemap.
Also useless: the growing pile of tools that claim access to Google’s internal AI ranking data. Google’s June 2026 guidance on third-party tools says plainly that none have it. Search Console’s Generative AI performance report is the source of Google data; Bing Webmaster Tools has an equivalent for Copilot.
How to measure it
- Search Console: Generative AI performance report, monthly
- Bing Webmaster Tools: AI Performance report and grounding queries
- Ask ChatGPT, Gemini and Perplexity your five main customer questions once a month; log who is named
- Call tracking on the phone number shown in your Business Profile
If you want us to run this for you, it is our AI SEO service, and it starts with a free audit that includes the AI answer check.
Want us to do this with you?
A free audit covers your listing, your site and what AI assistants say about your category. You keep the findings either way.
Call 1-800-481-8638 Request the free auditQuestions readers ask
Is AI SEO different from regular SEO?
No. Google states its AI features use the same ranking systems as search. AI SEO is regular SEO with more emphasis on direct answers, consistent facts and citations.
Do I need to block or allow AI crawlers?
Allow the retrieval crawlers (OAI-SearchBot, ChatGPT-User, Claude-SearchBot, PerplexityBot) if you want to appear in answers. Training crawlers are a separate choice.
Can a small local business show up in ChatGPT?
Yes. Local answers are decided by consistent business information, reviews and clear service pages, which a small operator controls.
Does llms.txt help with Google?
Google says it does not use it for Search. It may help other systems and is harmless if it returns 200 or a clean 404.
How long does AI visibility take?
Technical fixes are immediate. Citations build over months as content and mentions accumulate. Cited sources also turn over heavily month to month, so it is maintained, not achieved.