Here is the thing most people get wrong about AI search: they treat it as one channel. It is not. ChatGPT, Perplexity, and Google AI Overviews decide who to name using three genuinely different mechanisms, and a tactic that works beautifully on one can do almost nothing on another.
If your brand is missing from all three, the reason is usually different in each case. So let's take them one at a time.
ChatGPT: memory first, browsing second
ChatGPT answers from two places. Most of the time it draws on what the model already absorbed during training — a compressed impression of how often your brand appears alongside your category across the open web. When the question is time-sensitive or the model is unsure, it browses and pulls live pages.
That first mode is why brands with strong PR and community footprints get named in ChatGPT even when their own site is mediocre. You are competing on how widely and how consistently you are discussed elsewhere.
What to do:
- Get named on pages that already rank for "best [category]" lists. Roundups, comparison posts, and G2/Capterra-style aggregators are dense with brand-plus-category co-occurrence, which is exactly the pattern training absorbs. One inclusion in a widely syndicated roundup outperforms ten posts on your own blog.
- Standardise how you describe yourself, everywhere. One sentence, same wording, on your homepage, your LinkedIn page, your Crunchbase profile, every directory, every guest byline. Repetition of an identical descriptor is what makes a model confident enough to state it.
- Allow OAI-SearchBot and GPTBot in robots.txt. Plenty of sites blocked these during the scraping panic of 2023–24 and never revisited it. Check the file today; browsing mode cannot cite what it cannot fetch.
Perplexity: retrieval, and it reads the structure
Perplexity is closer to a search engine wearing a language model as a coat. Nearly every answer starts with a live retrieval pass, then it summarises the top documents and footnotes them. That makes it the most winnable of the three — and the fastest to respond to changes.
It rewards documents where the answer is findable in a single chunk. Long preamble hurts you here more than anywhere else.
What to do:
- Put the answer in the first 40 words under each H2. Question as the heading, direct answer immediately below, supporting detail after. Perplexity frequently quotes that opening sentence nearly verbatim.
- Publish the comparison table your competitors won't. Pricing, feature, and "X vs Y" tables get retrieved constantly because they answer multi-part questions in one block. Include the competitor names — that is the query text people actually type.
- Date and source your numbers. "As of Q2 2026" with a linked source beats an undated stat. Perplexity visibly favours documents that look current and verifiable.
Google AI Overviews: earn the ranking first
AI Overviews is the least mysterious of the three. It overwhelmingly synthesises from pages already performing in Google's organic index for that query, plus adjacent long-tail queries. If you are not in the top ten for the topic cluster, you are largely not in the running.
The good news: your existing SEO investment is the entry ticket. The work is making ranking pages quotable.
What to do:
- Target the question long-tail around the pages you already rank for. AI Overviews fires most often on informational and "how/why/can I" phrasing. Add those exact questions as H2s on the page that already has authority instead of launching a new one.
- Add FAQPage and Article JSON-LD to those pages. Structured data does not buy you a citation, but it makes the question-and-answer pairing unambiguous, and pages with it get pulled into Overviews at a noticeably higher rate.
- Keep the answer above the fold in the HTML. Content injected client-side or hidden behind accordions is inconsistently picked up. Render it server-side, in plain markup.
Same brand, three different jobs
Notice the pattern: ChatGPT is an off-site reputation problem, Perplexity is a content structure problem, and AI Overviews is a classic SEO problem with a formatting layer on top. Teams that pick one tactic and apply it everywhere end up visible on one surface and absent on the other two.
That is the work we do at Brightside GEO. We run all three under one strategy — directory and journal placements to build the co-occurrence ChatGPT relies on, content and schema restructuring for Perplexity's retrieval layer, and technical SEO to earn the rankings AI Overviews draws from — then track citation share monthly so you can see which surface is moving and which needs more weight.
Want to see where you currently stand on all three? Book a free audit →