AEO vs SEO: what changes, and how to split the budget

The honest size of the opportunity

Two numbers have to be held at once, and most writing on this subject holds only one.

AI search is small. AI referral traffic accounted for about 1.08% of all website traffic across industries in late 2025. For B2B tech firms it reached an average of 6.4% by January 2026. That is not a channel that replaces Google.

AI search converts about as well as organic, not several times better. The strongest evidence is a peer-reviewed study of 973 online shops and 164 million purchases (Kaiser and Schulze, Marketing Science, 2026), which found organic search converting about 13% better than ChatGPT referrals over 2024–25. Newer platform data leans the other way: Shopify found AI search visitors converting 49% better than organic in early 2026, and Adobe found AI visits to US retail sites converting 60% better than non-AI visits in July 2026. Those are the platforms’ own figures, and neither is anywhere near the large multiples sometimes quoted for this channel.

The fair reading is that an AI visitor is worth about the same as an organic one, with an edge on complex, high-consideration purchases. Much of AI’s value never shows up as a referral at all: a buyer who sees your name in an answer often searches for you directly, and arrives as branded or direct traffic.

Both facts hold. “AI search is tiny” and “AI search sends buyers at least as ready to act as organic search” describe the same channel.

The change that actually matters

Only 37.9% of cited URLs ranked in the first ten organic result blocks, down from approximately 76% in July 2025.

One caveat Ahrefs itself makes: it improved its parsing method between the two measurements, so part of that fall is method rather than market. BrightEdge, measuring differently, puts the top-ten share at about 17% and roughly flat across a year. The direction is not in doubt; the exact size is.

This is the finding that turns AEO from an SEO tactic into a separate discipline. A year ago, ranking well was most of the job — if you were on page one, you were in the citation pool. Now the majority of cited URLs are not page-one results at all.

It cuts both ways. Your page-one rankings no longer guarantee you get cited. And a competitor who cannot outrank you can still be the brand the assistant recommends, by being present in the sources the engine draws on.

How the disciplines differ

This page is about budget. For what the labels AEO, GEO and AI SEO mean and where they came from, see GEO vs AEO vs SEO.

	SEO	AEO
Goal	Rank your pages	Be named and cited in an answer
Main lever	Your own site and links to it	What third parties say about you
Unit of work	The page	The entity
Measurement	Position for a keyword	Citation share across a prompt set
Feedback speed	Days to weeks	Weeks to months, and noisy
Determinism	Broadly stable	Non-deterministic by design

The last row is the practical headache. The same prompt asked twice returns different sources, so AEO reporting has to be statistical where SEO reporting could be positional. Anyone showing you a clean AEO “rank” is smoothing over something.

A reasonable split

There is no universal ratio, but a defensible way to decide:

1. Measure your current AI referral share in analytics. If it is near zero and your category has little AI answer volume, AEO is a watching brief, not a budget line.
2. If it is growing month on month, fund AEO from the content portion of SEO spend rather than the technical portion. The two overlap most there, and content is where AEO’s specific requirements — extractability, statistics, ungated material — apply.
3. Fund earned media separately and honestly, as PR. It is the highest-impact AEO lever and it is not really SEO work at all.

If the question is whether to spend anything on AEO at all, rather than how to split it, is AEO worth it? works through the break-even maths.