Generative Engine Optimization (GEO): how to get AI to recommend your product (2026)
AI is the new search, and being recommended by ChatGPT or Perplexity is the new ranking. A solo, honest guide to Generative Engine Optimization (GEO) — what it is, the real numbers from a product I got AI to recommend from zero, and how to do it.
Solopreneur (20 years) · marketer & investor · 21 June 2026 · 6 min read
There’s a quiet shift happening under every solo founder’s feet: more and more people don’t search for a tool anymore — they ask an AI for one. “What’s a good free alternative to X?” And the model names three products. If yours isn’t one of them, you’re invisible to that customer — no ranking, no click, no second chance. Getting named in that answer is the new ranking, and the discipline for earning it has a name: Generative Engine Optimization (GEO).
Most solo advice treats AI search purely as a threat (“it’s killing my traffic” — and it is changing the game). This is the other half: AI is also a new, under-contested channel, and a one-person business can win it without a marketing team, a budget, or a single backlink. I know, because I did — here are the real numbers, then how.
The terms (so you can find the rest of the conversation)
It’s a young field with overlapping names — you’ll see all of these meaning roughly the same thing:
- GEO — Generative Engine Optimization (the most-used term)
- AEO — Answer Engine Optimization
- LLMO / LLM optimization
- AI SEO / AI visibility
The goal under all of them: be the answer the AI gives, not just a page in a list.
The real proof: a product AI recommends, from zero
Here’s a case from my own portfolio (kept anonymous on purpose). I launched a small global product — a one-person, vibecoded build — with zero backlinks and no ad budget. Within a few months, when I checked its analytics, one number jumped out:
That’s the whole thesis in one data point: an LLM decided this product was a good answer and started sending it qualified people — for free, repeatedly. That’s GEO working.
Why it happened (it’s not link-building)
The model recommended it for the same reasons a knowledgeable human would:
- Clear, narrow positioning — it was unambiguously “the X for Y”, so when someone asked an AI for exactly that, the match was obvious.
- It was genuinely a good answer — real, specific, useful — so being cited was deserved, not gamed.
- It was citable — clean, structured, easy for a model to read and extract, present in the kinds of pages models learn from.
Notice what’s not on that list: backlinks, domain age, ad spend. GEO rewards being the right, clear, readable answer — which is exactly the game a focused solo can win against bigger, vaguer competitors.
How to do GEO as a team of one
The free, do-it-yourself version of the method:
1. Target the questions people actually ask an AI
Not head keywords — questions and comparisons: “best [X] alternative”, “free [X] for [use-case]”, “[X] vs [Y]”. These are exactly what people type into ChatGPT, and what it answers with named recommendations. Position your pages (and your product) to be one of those names.
2. Be the answer — write to be extracted, not just read
LLMs lift clear, structured, specific passages: direct answers up top, comparison tables, tight lists, unambiguous claims. Vague, padded, “SEO-fluff” content is hard to quote — and uncited. Say the thing, plainly, where the model can grab it.
3. Be present where models source
Models learn from the open web — Reddit threads, comparison roundups, clear product pages, mentions. You don’t need a thousand links; you need to exist, clearly, in the places the answer is assembled from. Being genuinely useful in a niche discussion is GEO work.
4. The technical layer (cheap, future-friendly)
Clean, crawlable, structured content; schema; fast pages; and an llms.txt at your root — a
markdown map of your key content for AI (this site runs one). None of it is magic, but it’s low-effort
and it stacks.
5. Measure it — or you’re flying blind
You can’t improve what you can’t see, and GSC can’t see this at all (it’s Google Search only). Use
GA4: the “AI Assistant” channel, and Traffic acquisition by source/medium to spot
chatgpt.com, perplexity.ai, gemini.google.com. That’s how I caught the 6% — and how you’ll know
if it’s working for you.
The honest caveat (this matters)
GEO is young and less proven than SEO. The methods are still emerging, the models change, and the field is already crawling with hype-merchants selling “guaranteed AI rankings.” There’s no such thing. What I’m sharing is a real, measured result plus the emerging best practice that produced it — not a formula with a guarantee. Treat GEO as a cheap, compounding bet you measure honestly, and you’ll be ahead of 95% of people either ignoring it or being sold snake oil.
The solo + AI edge
Here’s why this is your game: the product above was built and optimised solo, by vibecoding it in an AI coding tool with my own prompts — and then the AI recommended it. The same skill that lets a one-person business build fast now also helps it get found. The tooling that levels you up is in the best AI tools for solopreneurs; the wider distribution picture is how to get traffic to a one-person business.
The takeaway
- Being recommended by AI is the new ranking — earn it with clear positioning + being citable, not links.
- It’s real and measurable: a solo product hit ~6% of traffic from ChatGPT, from zero, no backlinks.
- Target AI questions, be the extractable answer, be present where models source, run llms.txt, and measure in GA4 (GSC won’t show it).
- It’s young — measure honestly, don’t trust guarantees.
The era where you only optimised for ten blue links is closing. The solos who learn to be the answer — clearly, honestly, early — get a free, compounding channel the rest haven’t noticed yet.
Go deeper in this cluster
- GEO vs SEO: what changes when AI becomes the search engine — what’s the same, what’s new, and the order to do both in.
- How to track AI & ChatGPT traffic in GA4 — find your AI referrals (the “AI Assistant” channel) that Search Console can’t show you.
- How to get AI traffic to a new website — GEO from zero: no authority, no backlinks, fresh project.
- llms.txt explained — the robots.txt-for-AI, and how to write one.
The exact positioning and prompt workflow I used to get an AI to recommend a product from zero is the deeper playbook — more on that as this track grows.
Part of the complete SEO for solopreneurs guide.