The platforms
The agency’s FAQ answers the question directly. It optimizes for ChatGPT, Google Gemini and AI Overviews, Perplexity, Microsoft Copilot and Claude, and the voice assistants that draw on the same data, like Siri. Each uses different sources, so it covers the listings, directories and content each one relies on.
| Platform | What it draws on |
|---|---|
| ChatGPT | When it searches the web, composes an answer from the pages it retrieves and can recommend named local businesses |
| Google Gemini | Google’s index and Google Business Profile data |
| Google AI Overviews | The AI-written summary above Google’s regular results; a local business can be named inside it |
| Perplexity | Searches the web for every question and shows numbered source citations beside its answer |
| Claude | With web search enabled, reads retrieved pages and can recommend businesses by name |
| Microsoft Copilot | Draws heavily on Bing’s index and Bing Places data |
| Voice assistants such as Siri | Often answer from the same listing data the text assistants use |
How an answer is built
AI assistants don't run one search. They break a question into many smaller searches, check different sources for each, then combine the results. Bay Area AEO breaks the process into four steps. The assistant first works out what the person means. It then fans out, splitting the request into smaller searches: for a question about a pediatric dentist in Foster City, separate searches for pediatric dentists, for Foster City, for reviews and for insurance. It retrieves sources for each. Then it answers, combining what it found into a short recommendation.
Fan-out explains why the five signals matter together. Each smaller search can land on a different source: a listing, a review site, a service page, a local guide. A business that is consistent and specific across all of them survives every branch of the search and appears in the combined answer.
Why coverage has to be broad
A business that is strong on Google but missing from Bing may appear in Gemini’s answer and not in Copilot’s. A business with a clear website but thin reviews may be named by one assistant and described vaguely by another. The monthly run across every platform shows these gaps, and the cited-sources reading shows which listing or page to work on next.
AI SEO and Google’s AI features
The agency’s definition of AI SEO is SEO adapted for AI-powered search, including Google’s AI Overviews and AI Mode. Those two Google features matter because of reach: the agency estimates that about 60% of adults read Google’s AI Overviews. In the East Bay alone that is about 1,350,000 adults; in San Francisco about 438,000. An AI SEO plan that covers Google’s AI features alongside the chat assistants reaches customers wherever they ask.
Different assistants read different sources. Coverage means being right in all of them.
Spoken questions
Voice assistants change the shape of the question. The plumbing page puts it plainly: when a pipe bursts, nobody reads ten websites; they ask their phone who can come now. A spoken answer usually names fewer businesses than a written one, which makes accurate listing data even more important. Covering the voice assistants that draw on the same data, like Siri, means the same cleanup work reaches the spoken answer too.
Letting the assistants in
An assistant can only read a site it is allowed to fetch. The agency’s FAQ notes that if a site blocks AI crawlers it will say so and recommend the smallest fix. Opening the site to crawlers such as GPTBot, ClaudeBot or PerplexityBot, along with the search engines’ own crawlers, is a small step that lets every platform on the list read the business’s own words.