Here's what I do to show up in AI searches
With search now dominated by Google’s AI, I was worried about whether the previous SEO updates I did for my SaaS still work. So I went into this rabbit hole of researching between traditional SEO and how AI SEO works, and whether the usual SEO playbook I used still applies.
LLMs, especially Google’s, still use Google Search Rankings to find out what sites to suggest to people. So my work with backlinks, page speed, metadata optimization, and keyword research still helps build authority for my website. The better the authority is > the more Google trusts it > the higher the chance the AI would suggest it.
However, I also found that there are some more things I need to do in order to increase my likelihood of being recommended by AI. Here are the changes I’m planning to make:
Higher focus on category authority: Based on what I found, AI is more likely to recommend your product if you have content that positions your product as the best tool for a specific use. Having more people talk and recommend your product online is also helpful (testimonials, reviews on sites AI crawls on).
FAQs: AI seems to love FAQs, possibly because of the question - answer format it has. So making your FAQs have questions that people would most likely ask the AI would help.
Comparison content: The usual tool A vs tool B, or top X tools for… blogs also rank well for AI search. Put the features in bullets and write them in a conversational manner. I found that adding a few industry-specific keywords still works, but less than previously. A clear comparison table at the start also makes it AI-friendly.
Overall, I’m happy that I don’t have to start all over again with my SEO strategy. I just have to build on top of it in order to rank well on AI search.
____
PS: I started moving over to X/Twitter. I'll post more there in the future. Feel free to follow along if you're curious about me building in public.
https://x.com/Matt___Bauer
Great insights! I also think AI search is rewarding websites that demonstrate real expertise rather than just targeting keywords. I'm especially interested to see how much impact genuine user reviews and community mentions have compared to traditional backlinks over the next year.
The "AI SEO sits on top of classic SEO" framing matches what I bet my whole distribution on, so a data point from the other end of the spectrum: I run a catalog product (each catalog item is a name someone can lease), and instead of a handful of optimized pages I generate one static page per item — currently 162 of them, two languages, each with Product/Offer structured data and a genuinely specific title. The bet is exactly your category-authority point, just inverted: rather than being "the best tool for X," each page tries to be the single obvious answer to one hyper-specific query, because the buyer only searches in the rare week they need that exact thing.
Honest caveat since half the replies here are asking "any wins yet" — I launched last week, so I have zero ranking data and won't pretend otherwise. Ask me in two months. But two things already feel true from building it:
Gregory's point about third-party corroboration is the load-bearing one. My structured data and FAQs are just me talking about myself; what actually seems to move both Google and the AI overviews is other people mentioning the thing. The FAQ schema is table stakes, the off-site mentions are the moat — and those are the slow, unscalable part nobody wants to hear.
On schema vs. plain FAQ format (a few people asked): my weak prior is that the format matters more than the markup. The markup helps a crawler parse, but the LLM is reading rendered text anyway, so a clearly-written Q&A without schema probably beats sloppy Q&A with perfect JSON-LD. I'm carrying both and will try to isolate it later.
The part I have no answer for yet: does programmatic-SEO-at-scale (many thin-but-specific pages) read as authority or as spam to an AI ranker? That line feels like the whole game for anyone doing this, and I genuinely don't know which side I'm on until the data comes in.
This is quite helpful. I've just started and learning things from others' experience while implementing it, so i can iterate and test it fast has been fruitful for me. I've been thinking about the same thing and researched about it, read about the GEO and LLMO and having a clear sitemap.xml, while these works nice, the FAQ point you mentioned is really useful. Looking forward to learn more, if you've more to share from your journey.
Really interesting approach. I've been thinking about
this for my own site — a directory of AI tools for
content creators. The challenge is that AI search
results seem to favour established domains heavily
right now. Did you find that newer domains struggled
initially or did the structured content help overcome
that?
The category authority point is the one most people miss. You can have great backlinks and fast load times but if nobody is talking about your product outside your own site the AI has nothing to triangulate from. The comparison content thing tracks too. AI seems to pull heavily from pages that already have a clear winner framed in the answer which is why "best tool for X" style content performs so well even when the writing is average. Curious whether you've seen any difference between structured FAQ schema markup versus just writing FAQ style content without the markup. Wondering if the schema actually moves the needle or if it's the format itself that matters.
Some good insights here. I was talking to someone last week who talked about GEO and how he tries to get his brand mentioned in the AI summaries. Your point about FAQs directly relates to that as these can be specific quotable answers. He also talked about trying to have specific answers to the "people also ask" questions perhaps through blog posts on your website. Have you tried that at all?
This is useful. I’ve been thinking about this too because AI search feels like SEO plus positioning.
The part that stands out to me is category clarity. If a product is described too vaguely, AI probably won’t know when to recommend it, even if the product is good.
For my own SaaS, I’m realizing I need to be much more specific about the problem, the user, and the use case — not just “AI meeting notes,” but who it is for and when someone should use it.
FAQs and comparison pages make a lot of sense because they match how people actually ask AI tools for recommendations.
One thing to add from running this playbook at my own SaaS: your own site is only half the input, because models weight third-party corroboration heavily. Reviews, directory listings, comparison posts on other people's blogs, and forum threads mentioning your product move AI recommendations more than another FAQ page on your domain. Traditional SEO asked what your site says, AI search asks what the internet agrees on.
Great write-up. I'm curious whether you've noticed AI search favoring tools with interactive pages (like comparisons, benchmarks, or demos) over static blog posts. That's something I'm experimenting with for PromptProbe.
This is a good breakdown. I just went through the App Store version of this exact question — different platform, same underlying instinct: the algorithm cares less about keyword-stuffing and more about "does this thing have a clear, specific answer to a specific need." I rewrote my subtitle and keyword field this week to lead with one concrete problem (no-Content-ID background music) instead of generic terms like "AI" and "music" that were basically wasted since they're already in the app name.
The FAQ point especially maps over well — App Store search + Apple's own AI-ish recommendation surfaces seem to reward the same "be the obvious answer to one specific question" pattern rather than broad category terms. Comparison content is trickier on mobile since there's no real blog surface unless you build one on your own site.
Have you noticed AI search rewarding specificity over authority, or is authority still doing most of the work and specificity is just a tiebreaker?
I like how you framed this as “AI SEO is on top of normal SEO, not instead of it.” Feels like a lot of founders are throwing away backlinks/technical basics and praying to the LLM gods. Category authority + FAQs + comparison pages is basically “be the best answer on the internet,” which is exactly what AI wants to plagiarize.
Curious if you’ve seen any clear wins yet from tweaking FAQs/comparison posts specifically for AI answers?