Why AEO won't fix your early-stage growth

AI visibility is being sold as the magic pill for growth in 2026. I spent the last year running SEO and AEO programs for companies like Lovable, Webflow, Surge AI and Augment Code, and teaching workshops on how to do it.

So my answer might surprise you. When people ask "Should we prioritize SEO or AEO right now?", I usually say:

Neither.

I'm a big believer that for most early stage startups, one channel ends up driving about 80% of the growth. If you haven't found PMF yet, and you haven't found that one "double-down" channel, SEO and AEO are almost always the wrong place to focus.

Why? Well, the channel takes months to start working, it converts lower on average, and it doesn't do much unless you can feed it from a go-to-market that's already converting.

This post makes the case for waiting. It also covers how I'd go find the channel to bet on in the meantime. Already past that bar? The companion piece is the one for you: the plan I'd run to build an AEO program.

Quick primer: What is SEO and AEO

SEO, or search engine optimization, is about structuring your website so that Google ranks your pages near the top when somebody searches a relevant query. Links, content, technical hygiene, topical authority.

AEO, or answer engine optimization, is that same job for AI assistants. Somebody asks ChatGPT, Claude, Perplexity, or Gemini a question, and the model picks which sources get read, cited, or paraphrased into the answer. AEO is the work of becoming one of the sources it picks.

Why do people invest in SEO and AEO?

A single keyword can be worth $10,000+ a month, and a family of related prompts (the AEO equivalent) is worth even more than that.

Take "best corporate credit card for small businesses". The phrase itself does 600 US searches a month, and its parent topic, "best business credit cards", does 37,000/mo. Now look at page one for it. It's a competitive page. NerdWallet holds the top organic slot. Brex's product page is buried at position 10. Ramp and Capital One are paying for the ads at the bottom.

For a self-serve product like Ramp or Brex, a top ranking pays out a few ways.

The obvious one is traffic. NerdWallet's #1 article pulls ~45,000 monthly visits across hundreds of related queries. Buying that through Google Ads would run you ~$550K/mo, and some slice of those visits becomes self-serve signups on top.

The less obvious one, the brands ranking organically tend to be the ones LLMs cite. You do the SEO work and the AEO answer comes with it. Ask Claude for the best small-business card and it gives you Brex, Ramp, Chase Ink, Amex. The same page-one names (I talk about our work with Ramp at GrowthX here btw).

And this is one keyword, remember. A real program targets hundreds of them. Every "best X for Y". Every "alternatives to". Every category, every competitor. Nobody's going to hand you a clean ROI number on this, same as nobody hands you one for a brand survey, but influencing what the models say is worth something real. The question is when to invest.

So why not invest?

Organic is slow, but you need to learn fast

If you don't have strong PMF, your constraint as a founder is time - how fast you can iterate the product. So founder sales, warm intros, events, direct outreach. Those work best here. None of these scale incredibly well, but you get unfiltered access to your buyers out of them.

Organic content is too slow and too detached to help here. A new site has zero authority, so content takes 3-6 months to even rank. That's 3-6 months of guessing at what your buyer wants instead of hearing it from them. Like stopping mid-hike to build a suspension bridge over a pond you could wade across in a minute.

I got to experience this first-hand. At Kite we had 5 million Python developers a month on the docs site and tens of thousands of signups. None of it moved the company forward in any meaningful way. Traffic can't fix a product that's not ready.

You must find the double-down channel

Early on, growth tends to follow an 80/20 rule. One channel ends up driving most of it.

GrowthX was like this. Most of our pipeline came from events, physical ones and virtual ones. We sold a new category, so the job was double, educating (2h workshops) and building trust (fancy dinners in SF) at the same time. SEO, for us at that point, would have been a waste of time.

More often than not, you find these "double-down" channels by pushing hard on your unfair advantages. A founder who can sell, a network, a wedge audience that trusts you. Organic content almost never wins that race from a standing start. Amplifying, that's where it helps - the channel's already converting, and the sales calls and customer wins it produces get turned into pages that compound.

Here's how that priority shifts by stage for a sales-driven B2B SaaS business:

As for finding the channel, here's my rough order.

1. Write down your unfair advantages. A founder who can sell, a network you can mobilize, a wedge audience that already trusts you, a distribution edge. The double-down channel tends to come out of one of these, and you won't find it on a best-practices list.
2. Run small, real tests on the two or three channels that teach you fastest. Founder sales, warm intros, events, direct outreach, that kind of thing. Three answers you want before any money goes in: who buys and the job they're hiring you for, which channel teaches you fastest for real, and whether its attention turns into pipeline you can predict.
3. Watch for the channel that converts and that you can push harder on without it breaking. That one's your double-down candidate. Converts-but-caps-at-ten-conversations-a-week is a different bet than a channel you can keep feeding.
4. When you find it, concentrate there until it's clearly working before opening a second channel. The usual failure, five half-channels because one of them twitched. A little SEO, a little founder social, a little outbound, a few events, no real force behind any of it.
5. Until then, only publish artifacts from what you're learning. A product page after ten sales calls, a customer story after one surprises you. The site's getting stronger because the positioning is getting codified, and rankings have nothing to do with it (yet).

AEO changed the math, so "wait" doesn't mean "ignore"

With all that said, ChatGPT and Claude really have increased how much content influences buying decisions.

The behavior shift, you can see it in yourself. I use Google directly maybe a fifth as much as I used to. One product-evaluation prompt fires off ten or twenty searches, reads more pages than I'd ever open, and hands the answer back. A lot of the time the buyer never clicks through to anything.

Trust laundering is the bigger shift, though. There's audiences that never trusted a random Google result, enterprise procurement, developers, and they're getting the same information synthesized inside Claude or ChatGPT now. The source gets obfuscated, the model ends up holding the trust. So AEO becomes brand work aimed at a new set of buyers - Claude, ChatGPT, and Perplexity deciding who gets cited, and how somebody hears about you the first time.

The numbers, they're showing it already. Webflow gets about 8% of self-serve signups from LLMs now, and those convert at six times the rate of non-branded organic search. Adobe Analytics measured AI-driven traffic to US retail sites up 693% year-over-year over the 2025 holidays, converting 31% better than other sources too.

So no, "not yet" doesn't mean "never". Find the channel, get it compounding, and then AEO deserves real investment. Earlier than most people think, probably.

What's next

Say you're past the bar. You know who buys, a channel works, there's source material to pull from. What's left is building the program, and that's its own post: the plan I'd run to build an AEO program. It's the two things that stay true no matter how the tactics change, plus the order I'd do the work in.

If you want to talk any of this through for your company, reach out.