what is geo/ai search, and why should you know about it? (save this) some developer tools are getting 80% of their customers from llms. how does this organic traffic happen? how does it compare to seo? this is our explainer video, hope it helps anyone!

i reverse engineered how supabase became the #1# recommended backend in AI search and where they're still losing 37.5% citation share. 99k github stars. $5B valuation. the default database for every vibe coding tool on the internet. when u ask chatgpt "firebase alternative with a SQL database" - supabase gets the blue link. "best managed postgres for a startup" - supabase. "add a database, auth, and storage to my react app without building a backend" - supabase. heres how they got there: > they built on postgres instead of inventing their own database. 30 years of trust, borrowed overnight. every developer already knows SQL. > they wrote comparison pages on their own site. "supabase vs firebase." "supabase vs auth0." "supabase vs heroku postgres." when someone asks the AI which backend to use, supabase has a page ready to be cited. > they invented "launch week" - every quarter they ship 5 major features in 5 days. 15 launch weeks so far. thats 75+ individual moments that each generate blog posts, backlinks, and content the AI trains on. > they shipped pgvector in early 2023, six months before most devs knew what a vector database was. by the time the AI wave hit, supabase was already the default answer to "where do i store embeddings." > they became the "connect to supabase" button in lovable, bolt, replit, and cursor. not because of deals. because supabase was already everywhere in the training data. but heres what blew my mind: supabase gets 0 citations on chatgpt for queries about their own core features: > "how do i make sure each user can only read their own rows in postgres" - this is literally row level security. supabase's signature feature. chatgpt doesnt cite them. doesnt even mention them. > "how do i push realtime updates to the browser when my database changes" - realtime is one of their 5 core products. 0 citations. > "how do i add google login and magic links to a next.js app" - supabase auth does exactly this. 0 citations. > "best vector database for a RAG app" - they shipped pgvector before almost everyone. 0 citations. qdrant and aws get cited instead. the pattern: supabase wins every query where someone names the category. but when someone describes the job they need done (the exact problems supabase solves) the AI has no page to link to. firebase is even wilder. mentioned 15 times across all queries but only cited twice. every AI knows firebase exists. none of them have a reason to link to it. supabase built one of the greatest developer platforms ever. and even they have blind spots in AI search. imagine what ur company is missing.

ppl really be spending thousands on ads, emails and inboxes, launch videos and so on but wont spend a couple of hours each week to make sure that llm's recommend them when a potential user asks its literally free traffic and free users but everyone ive talked to puts it off cuz its a black box and it feels not as pressing this could be you 🫶 lock in

dm me ur company url and i'll tell u exactly what chatgpt recommends when someone searches for a tool in ur category takes me 30 minutes ill make you a custom dashboard that looks like this :) it'll show you: > the queries where u get cited > the queries where ur competitor gets cited instead > what type of content and pages mainly get cited > where llm's go wrong when describing your product > a content plan for how to fix all these issues > and more this is pretty expensive to run so not everyone will get an audit. first come first serve :)

we've audited AI search visibility for 10 devtools in the last month every company that ranks #1# in their category does the same 4 things: > they have a page for every "how do i [problem]" query in their space. not feature pages. pages that answer the question first and mention the product second. > they have comparison pages for every competitor. vercel has 13 "[vercel] vs [competitor]" pages. when someone asks chatgpt "vercel vs netlify" theres a page ready to be cited. > they answer cost queries. "cheapest [category] for startups". most companies ignore this. elevenlabs loses every single cost query to amazon polly and deepgram because they have no page addressing price. > they write pages matching the exact phrasing ppl type into AI. not "our analytics platform" but "how to measure user onboarding drop off." amplitude has hundreds of these. posthog has zero. companies doing all 4: vercel (48% citation share), elevenlabs (33%), zapier (25%) companies missing 2+: losing to competitors 1/10th their size dm me if u run a devtool and want to know where u stand we'll run the audit free

wow got a lot of ppl wanting audits from this haha itll take a bit but dont worry we'll get everyone a deep audit on where their product stands in getting recommended by AI :)

talked to a devtool founder doing $2M ARR who gets zero traffic from AI search meanwhile a competitor with 1/10th their users is getting cited by chatgpt on every query in their category the difference came down to one thing: the competitor wrote pages titled exactly like the questions ppl ask before they know the product exists > "how do i do X without Y" > "best tool for [use case] under $50/mo" > "X alternative that works with [integration]" the founder had better docs, more github stars, bigger brand but when someone asked chatgpt for a recommendation, there was nothing to link to we ran the audit in 30 minutes. found 14 queries where competitors were getting cited and they weren't. they wrote 6 pages in a week. already showing up on 3 of them. dm me if u want me to run this on ur company takes 30 min and its free

we reverse engineered how zapier became the #1# recommended automation tool in AI search 22% citation share. more than n8n and make combined. everyone talks about their google SEO (9M+ monthly visits). but heres what actually makes them win in AI search: > a dedicated page for every app integration combo. "connect gmail to slack." "connect stripe to airtable." 8,000+ apps means thousands of pages. when someone asks chatgpt "how do i automatically send form submissions to a google sheet and slack" zapier has an exact page for that. > they were first to frame automation as AI agents. zapier now leads with "AI agents and chatbots" not just "connect apps." when someone asks "tool to build AI agents that can use my existing apps" zapier gets cited #1#. > comparison pages they wrote first and control. "zapier vs make vs n8n" — they wrote it, they rank, they get cited on both chatgpt and gemini. > they turned integrations into a content moat. every new app that joins = hundreds of new pages = hundreds of new queries they can get cited on. chatgpt cited them on 6 out of 18 prompts. gemini on 3 out of 17. the playbook: they didnt just build the most integrations. they made sure every integration had a page the AI could cite. 8,000 integrations = 8,000 answers to questions ppl ask chatgpt.

talked to a devtool founder last week who couldn't figure out why signups dropped 30% in 3 months their traffic was the same. pricing didn't change. product got better. but his competitor started showing up in chatgpt answers for every query in his category > "best [category] tool for startups" > "how do i [problem] without [pain point]" > "[his product] alternatives that are cheaper" we looked at the actual citation data: competitor: cited on 12 out of 18 queries his product: cited on 2 that really hurts haha every day someone asks AI for a recommendation in his category, and every day the AI sends them somewhere else. he never would have known if he hadn't checked have u checked?

elevenlabs is the #1# AI voice company in the world and they're getting 0 citations on chatgpt for "best AI text to speech that sounds human" think about that. the most obvious query for their product. and chatgpt doesn't link to them. we ran 18 queries across chatgpt and gemini: what they win: > "best AI voice cloning tool" -> cited #1# > "how do i dub a youtube video into spanish" -> cited #1# > "AI voice with emotional control" -> cited #1# > "best text to speech API for developers" -> cited #1# what they lose: > "cheapest realistic tts API" goes to amazon polly, deepgram, google > "elevenlabs alternatives that are cheaper" deepgram gets cited, elevenlabs doesn't > "best conversational voice AI for customer support" goes to retell, vapi, bland > "best AI text to speech that sounds human" 0 citations on chatgpt elevenlabs owns quality. they lose every single cost query. anyone searching "cheapest tts API" has budget and is ready to buy. and right now amazon polly and deepgram are getting that click. the AI recommends whoever wrote the best answer to the question. not necessarily whoever has the best product.

we been talking to geo/aeo people at some of the top devtools and the same thing keeps happening everyone is figuring this out alone there isnt any one size fits all playbook. its kinda a black box. the most common question i get from every single one of them: "what are the other companies doing?" thinking about starting a small private group for ppl actually doing this work. but wanna gauge interest first. most of these ppl already wanna join if thats u or someone u know, lmk :)

in my opinion if i was a growth stage devtool founder and choosing an ai search tool i’d probably look for something that: a. not only tracks analytics but also proactively does the practical tasks required for more visibility (bc they should be able to in 2026) b. the team on that tool understands my product and its landspace on geo specifically and there’s quantitative data i can see on that c. they move fast and are communicative, as in i know exactly what the tool is doing week by week (also quantitative) d. oh i’d also probably want part of the deal to be performance based as well

"what even is GEO and how do i do it?" GEO (generative engine optimization): >search ur product category in chatgpt, perplexity, and claude right now. if ur not in the answer dont worry its fixable >comparison posts are the #1# cited content format. write "[ur product] vs [competitor]" for ur top 5 competitors this month. >reddit mentions with upvotes are weighted heavily. show up on reddit weekly. (tho this might be changing soon haha) >directory listings are trust signals. submit to every niche directory in ur category today. takes 5 min each. >structured Q&A content gets cited more than narratives. write posts that directly answer "what is the best tool for [X]" >ur docs and readme are fair game. write them in natural language not jargon. >freshness matters. a post from this month outranks a post from last year. (make sure to mark the last edited date of your content) >perplexity and chatgpt pull from different sources. check both. >google rankings and AI citations are different games. u can rank #1# on google and not be number 1 recommended in chatgpt. >AI models prefer third party mentions over first party content. get other ppl talking about u. >check ur GEO rankings weekly. this is a moving target not a set-and-forget. the founders doing this right now will own their category in AI search by 2027.

the cold email deliverability checklist: >separate domain for outbound. never send from ur main domain. ever. >warm up the domain for 2 weeks minimum before sending a single cold email >plain text only. no images. no html signatures. no tracking pixels. >no link in the first email. the first email is supposed to be a hook >30 sends per mailbox per day. not 50. 30. >rotate 3-5 mailboxes so no single one gets burned >verify every email at send time not at import. data decays daily. >follow up 3 times minimum. over 50% of positive replies come from follow-ups 2-4. >if ur reply rate is below 3% the list is bad. stop blaming the copy.

the most slept on outbound strategy in 2026: >go to linkedin and search for companies hiring for the role ur product replaces >a company posting "hiring a social media manager" is a company that needs social media help RIGHT NOW >thats a warm lead. they just told u their budget and their pain. now: >pull the hiring managers name from the job post >enrich their email using apollo, prospeo, or any waterfall provider (like jam ;) >check their company linkedin and website to understand what they actually sell >write a 3 line email: reference the job post, explain how ur product does what theyre hiring for at a fraction of the cost, ask for 15 min thats it. u are emailing someone who literally just publicly announced they have the problem u solve and they have budget allocated for it.