Ask a marketing team where their brand "shows up in AI" and you will usually get one confident answer. That confidence is often misplaced. Google's AI Overviews and standalone AI chat assistants like ChatGPT, Claude, and Perplexity are not the same surface, and treating them as one metric is the fastest way to misread your own visibility.
Both use large language models. Both produce a synthesized answer instead of ten blue links. Past that, the similarities mostly stop. One is an extension of the search engine you have optimized for since 2005. The others are newer retrieval systems with their own rules, and in some cases only a loose relationship to your Google ranking at all.
This matters because brands increasingly win one surface and lose the other, and never notice, because they are only measuring one of them.
AI Overviews sit inside Google Search, above or alongside the organic results. Google has been explicit about the mechanics: the feature draws on the same retrieval and ranking systems that power classic search, then adds a generation layer on top. In its own guide to optimizing for generative AI features, Google states plainly that "the best practices for SEO continue to be relevant because our generative AI features on Google Search are rooted in our core Search ranking and quality systems."
Google also describes a "query fan-out" process, where a single search expands into several related queries behind the scenes so the system can pull a wider set of pages before writing the summary. That is a bigger net, not a different lake. The pages being pulled are still pages Google's index already trusts.
Practically, this means an AI Overview behaves closer to a rich answer box than to a chatbot response. If your page ranks, is crawlable, and answers the query cleanly, you are already in the pool of candidates. Google's own documentation on AI features is direct about this: there are no additional requirements to appear in AI Overviews beyond the standard requirements for appearing in Search at all. We covered this relationship in more depth in AI search vs Google, but the short version holds: AI Overviews are search-adjacent, not search-independent.
Structured data still plays a supporting role. Clear headings, a direct answer stated plainly near the top of a page, and schema markup that describes what the page is about all make it easier for Google's retrieval layer to lift an accurate excerpt into the summary. None of this is new advice. It is the same guidance that has applied to featured snippets for years, now feeding a different piece of the results page.
ChatGPT, Claude, and Perplexity did not inherit a search index built over two decades. Each has assembled its own view of the web, blended with model training data, and each retrieves and weighs sources differently.
Perplexity behaves the most like a search engine: it tends to run a live web query for most prompts and builds its answer from freshly retrieved pages, with inline citations tied closely to specific claims. ChatGPT leans more on a mix of what the model already learned during training and a narrower, more selective set of live retrievals, which means a page can shape an answer without ever being named as a source. Claude and Copilot sit somewhere between those two patterns, depending on whether web search is active for a given query.
None of this runs through Google's ranking systems. A domain with strong backlinks and a page-one ranking for its target keyword can be functionally invisible inside a chat assistant's answer, while a smaller site with a well-structured comparison page and consistent third-party mentions gets cited instead. This is the core idea behind generative engine optimization, which we unpack in what is GEO.
Gemini and Copilot complicate the picture further, since each blends its own assistant behavior with a search backend it partly controls. Gemini can draw on Google's index directly, while Copilot leans on Bing. Training data cutoffs add another wrinkle: a chat assistant answering from memory rather than live retrieval may be describing a version of your product, your pricing, or your positioning that is months out of date, even when your website has already moved on.
Because the inputs differ, the outputs diverge, sometimes sharply.
A brand with a decade of SEO investment, strong domain authority, and a well-optimized product page is well positioned for AI Overviews. Google's summary layer reads from the same index that already ranks that page. If you show up in the top results, you have a real shot at being one of the sources cited in the overview.
That same brand can be nearly absent from a ChatGPT or Perplexity answer to an adjacent question, like "what is the best tool for X" or "how does pricing usually work in this category." Chat assistants tend to reward brands that are discussed in places an LLM actually reads during training or retrieval: comparison articles, review sites, forums, documentation, and other companies' blog posts that mention them by name. A page can be perfectly optimized for Google and still never surface in a chat answer, because it was never picked up anywhere an LLM treats as a citable source.
The reverse also happens. A newer brand with little domain authority but an active presence in relevant online discussions, structured comparison content, and consistent third-party mentions can be recommended by AI chat assistants well before it ever ranks on page one of Google.
Picture two project management tools competing for the same buyer. Tool A has ten years of content marketing, hundreds of backlinks, and a page-one ranking for "best project management software." Tool B launched two years ago with a thinner backlink profile, but its founder is active in relevant online communities, its comparison pages are frequently linked from independent review sites, and several unrelated blogs mention it by name alongside larger competitors.
Search Google for "best project management software" and Tool A is very likely inside the AI Overview, because Google's index already trusts the pages that already rank. Ask ChatGPT or Perplexity the same question and Tool B has a real chance of being named, because those systems weigh the surrounding conversation about a brand more heavily than the brand's own domain authority.
Neither result is wrong. They are answers to the same question, generated by systems that trust different evidence.
Teams that check only one surface tend to assume the other looks roughly the same. In practice, the set of sources an AI Overview cites and the set of sources a chat assistant cites for a comparable question often only partially intersect.
That follows naturally from the retrieval differences above: one system reads a search index built over two decades, the others read a mix of training data and live web pages selected by different heuristics entirely. There is no shared source of truth to check against, and no single dashboard that tells you how your brand appears across both by default.
This is exactly the gap that turns AI visibility into an ongoing monitoring problem rather than a one-time optimization project. You cannot infer your Perplexity presence from Google Search Console, and you cannot infer your AI Overview presence from a handful of prompts typed manually into ChatGPT.
If AI Overviews and AI chat assistants pull from different systems, then "are we visible in AI" is not one question. It is at least two, and arguably five, once Gemini and Copilot are counted as their own surfaces with their own quirks.
Tracking needs to reflect that split. A useful setup checks:
This is the entire premise behind Bold GEO: a daily refresh that checks how your brand is cited across ChatGPT, Perplexity, Gemini, Claude, and Copilot, tracked as separate surfaces instead of one blended score.
It depends on where your buyers actually start looking.
If your category still starts with a Google search, a mix of branded and category queries, AI Overviews deserve first attention, and the work looks a lot like SEO: technical health, page structure, and content that answers the query directly. Your existing SEO team can likely extend into this without a new playbook.
If your buyers increasingly ask ChatGPT or Perplexity directly for recommendations before they ever type a query into Google, chat visibility needs its own workstream: getting mentioned accurately on the sites and forums these models draw from, keeping your product facts consistent everywhere they appear, and showing up in the comparison content your category already searches for.
Most B2B categories sit somewhere in between right now, which is exactly why treating this as one undifferentiated "AI visibility" number under-serves both efforts.
A workable approach separates the tracking from the tactics, but keeps both under one roof.
Start by auditing where you currently stand on each surface for a representative set of queries: branded, category, and comparison. Note not just whether you appear, but which pages or sources get cited when you do. That tells you whether the gap is a content gap, a technical gap, or a reputation gap.
From there, treat AI Overview visibility as an extension of your SEO roadmap. Treat chat visibility as its own line item: earned mentions, structured facts, and a presence in the third-party content these models already trust. Review both on a recurring cadence rather than a one-time check, since model behavior and search results both move continuously. You can start monitoring both surfaces from a single Bold GEO account rather than piecing together two separate workflows.
Some of the highest-leverage work for chat visibility never touches your own website. It means making sure your pricing, positioning, and category claims read the same way on your site, your review site listings, your documentation, and any third-party comparison content you can influence. Inconsistent facts across sources give a model less basis for treating any single claim as reliable, which makes it less likely to cite you with confidence.
It also helps to think in terms of surfaces rather than keywords. A page written to answer one specific question in a direct, well-structured way tends to travel better across both AI Overviews and chat assistants than a page written primarily to rank for a broad head term. Specificity reads as trustworthy to a retrieval system, human or otherwise.
The brands that get this right are not the ones that "do AI SEO." They are the ones that stopped assuming a single checklist could cover two structurally different systems.
Bold GEO monitors how your brand is cited across ChatGPT, Perplexity, Gemini, Claude, and Copilot on a daily refresh. 7-day free trial, no credit card.