Skip to content
Waseit
← All articles
AEOGEO

The visitor AI sends is worth more, and SEO is being bought

In one year, the traffic AI assistants send to US retail sites went from less valuable to more valuable than the rest. Meanwhile Adobe bought Semrush for 1.9 billion dollars to sell visibility inside AI answers.

42% better

conversion of AI-referred traffic against the rest on US retail sites, March 2026 (Adobe, reported by TechCrunch)

September 22, 20268 min read
In this article
  1. In one year, from worse to better
  2. How many are they? Nobody says
  3. The search optimisation industry is being bought
  4. The weak point sits on the website
  5. And for a local business
  6. What this reading does not prove
  7. What it changes

Short answer: two signals arrived in the same season and they say the same thing. The visitor an AI assistant sends to a retail site now buys more than the others, and the money of the search optimisation industry has reorganised around that visitor. The first shows in Adobe's traffic data, the second in a 1.9 billion dollar acquisition. Neither says how many of these visitors there are, and that is the question to keep in mind.

1.9 billion dollars

value of Adobe's acquisition of Semrush, announced 19 November 2025

plus 1,324%

AI traffic to US retail sites between October 2024 and May 2026 (Adobe)

A visitor who arrives after talking to an AI has often already compared. They are not coming to look, they are coming to buy.

In one year, from worse to better

According to Adobe's quarterly report, as reported by TechCrunch on 16 April 2026 and built on more than a trillion visits to US retail sites, traffic from AI assistants converted 42% better than the rest in March 2026. A year earlier, in March 2025, it converted 38% worse. Revenue per visit made the same swing: 37% above ordinary traffic in March 2026, when twelve months earlier the ordinary visitor was still worth 128% more.

The simplest explanation, and it is an interpretation, not a measurement: the conversation with the assistant does part of the salesperson's work. The customer compares, asks questions, rules options out, then clicks. They land on the site with the decision half made.

How many are they? Nobody says

Adobe publishes striking growth rates: plus 269% year on year in March 2026 according to its 20 April release, plus 1,324% between October 2024 and May 2026 according to its 17 June release, and plus 2,215% for travel over the same period. None of these releases gives the share of this traffic in the total. A growth rate says nothing about size: multiplying a small thing by fourteen can still give a small thing. That is the caveat missing from most articles that repeat these figures.

The search optimisation industry is being bought

On 19 November 2025 Adobe announced the acquisition of Semrush at 12 dollars a share, about 1.9 billion dollars. Semrush is one of the most widely used search optimisation tools in the world. Yet Adobe does not present it as an SEO tool but as a brand visibility platform providing generative engine optimisation (GEO) and SEO solutions, and justifies the deal by one need: remaining discoverable in AI search. The acquisition closed at the end of April 2026, and on 17 June Adobe launched a product, Adobe Brand Visibility, combining Semrush's AI visibility data with its own content tools.

That is the economic signal of the season: a marketing software vendor is betting nearly two billion on the idea that visibility inside AI answers has become a budget line of its own. It must also be read the other way round: whoever publishes the AI traffic growth figures is whoever sells the remedy. That does not make the data false, but it does not make it neutral.

The weak point sits on the website

The same report, again according to TechCrunch, adds a less discussed finding: around 34% of product pages on the retail sites studied cannot be properly read by an AI, and about a quarter of homepage content is not prepared for it. In other words, the most valuable visitor of the moment arrives through a channel that misreads one page in three.

Read on 21 September 2026. Adobe's releases were read at source. Lines marked relayed come from Adobe's quarterly report as cited by TechCrunch: the report itself did not respond to our three attempts to read it.
SignalMeasurementPeriodSource
AI traffic to US retail sitesplus 269% year on yearMarch 2026Adobe, release of 20 April 2026
The same traffic, over nineteen monthsplus 1,324%October 2024 to May 2026Adobe, release of 17 June 2026
AI traffic to travel sitesplus 2,215%October 2024 to May 2026Adobe, release of 17 June 2026
Conversion of AI traffic against the rest42% better, against 38% worse a year earlierMarch 2026Adobe, relayed by TechCrunch
Revenue per visit from AI traffic37% above the restMarch 2026Adobe, relayed by TechCrunch
Adobe's acquisition of Semrushabout 1.9 billion dollarsannounced 19 November 2025Adobe, release

And for a local business

These data cover online retail, not the tradesperson or the neighbourhood practice, and they should not be transposed as they are. But the mechanism does transpose. For a local business, conversion is not a completed basket, it is a call, a quote request, a visit. If the assistant does the comparing beforehand, the contacts that arrive are fewer and better qualified, and the ones that do not arrive were lost inside an answer the business never saw.

What this reading does not prove

  • That AI traffic weighs a lot.

    None of the sources cited gives its share of the total. Growth is strong, size remains unknown to the reader.

  • That we verified the conversion figure.

    It comes from Adobe's quarterly report as cited by TechCrunch; the report did not respond to our attempts to read it, and we say so rather than stay silent.

  • That the measurement is neutral.

    Adobe now sells an AI visibility product. Its traffic figures also serve its sales pitch.

  • That online retail tells the truth about local services.

    A basket and a quote do not behave the same way. Transposing to local is reasoning, not measurement.

What it changes

Count contacts, not only visits.

If AI does the sorting beforehand, volume falls and quality rises. A dashboard that only watches traffic will read good news as bad news.

Make your pages machine-readable.

One page in three misread on well-equipped retail sites: a local site with no technical budget is likely to do worse.

Distrust figures without a denominator.

A plus 1,324% without the share of the total justifies no budget on its own.

Know what AI says about you before paying someone to change it.

The big vendors now sell visibility inside AI. The first step, reading what assistants answer about your trade and your city, costs nothing.

Frequently asked questions

Does traffic sent by ChatGPT or Gemini really convert better?
On US retail sites, yes according to Adobe: in March 2026 this traffic converted 42% better than the rest, against 38% worse a year earlier, according to its quarterly report cited by TechCrunch on 16 April 2026. We could not read the report itself, and the figure says nothing about this traffic's share of the total.
Why did Adobe buy Semrush?
Adobe explains it in its release of 19 November 2025: Semrush is described there as a brand visibility platform doing GEO and SEO, and the deal answers the need to remain discoverable in AI search. The acquisition is worth about 1.9 billion dollars and closed at the end of April 2026.
What does this change for a small local business?
Fewer contacts, better prepared, and part of the choice made inside an answer you do not see. The practical response is to read what assistants say about your trade in your city and make your site machine-readable, which needs neither a paid tool nor a provider.