Adobe now sells three acronyms in one product line. SEO, which has been around since the nineties. GEO, generative engine optimization, which means being named inside the answer the model writes. And ASO, agentic search optimization, for the case where the thing doing the searching is a piece of software buying on someone’s behalf.

Eighteen months ago two of those three did not exist as product categories.

They exist now because Adobe closed its acquisition of Semrush on April 28 for $1.87 billion in cash, and because two months earlier Profound raised $96 million at a $1 billion valuation, led by Lightspeed with Sequoia and Kleiner Perkins alongside.

A category with two funding events that size in one half-year is a category that will land on your budget next quarter, whether or not you asked for it.

Read the 269% carefully

Semrush brought $480 million of annual recurring revenue with it, which is a real number but not the reason for the price.

The reason is in Adobe’s own measurement. AI-referred traffic to US retail sites grew 269% year over year in March 2026.

Anil Chakravarthy, who runs Adobe’s Digital Experience business, put the thesis plainly when the deal was announced: “Brand visibility is being reshaped by generative AI, and brands that don’t embrace this new opportunity risk losing relevance and revenue.”

Now the caveat, because that 269% gets quoted as though it settles something. It is growth off a small base, on US retail. It does not say that a B2B brand in Central America is receiving meaningful traffic from ChatGPT today. It says the curve points up, and that Adobe was willing to pay $1.87 billion on the assumption it keeps pointing that way.

Those are different claims and marketing teams keep collapsing them into one.

ASO is the one worth watching, and almost nobody is set up for it. Agentic search optimization assumes the thing evaluating your company is software running a purchase on someone’s behalf, comparing you against three competitors against criteria it was handed, and never rendering your page. No brand impression, no visual identity, no landing page. Whatever your site says in machine-readable form is your entire pitch. Adobe put that acronym on a price list before most companies have written a single structured fact about themselves.

A dashboard tells you that, not why

Profound tracks mentions, sentiment and performance across answer engines, and its customer list runs to Target, Walmart, Figma, MongoDB, US Bank and Charlotte Tilbury. Large brands with staff assigned to this full time.

Buying the same tool does not buy you their result.

Two failure modes show up immediately. The first is sampling noise. Identical prompts return different sources, so a dashboard can show a clean upward line that is entirely variance, and somebody moves budget because of it. We went through the arithmetic in AI visibility measurement and sampling noise.

The second is assuming press coverage transfers. In Muck Rack’s December 2025 measurement, the journalists PR teams pitch and the ones AI ends up citing overlapped by just 2%. Getting into the right publication and getting cited by the model are close to unrelated games.

Build the baseline before you buy anything

Write down ten to fifteen questions a real buyer in your category would ask a model today. Full questions with the context that person would include, not keywords. “Which company should I hire for X if I run a Y business in Costa Rica?” is a question. “X company” is a keyword and it will not tell you anything here.

Then ask them yourself. Two different models, three runs each, this week. Record whether you appear, where in the answer, and how you get described.

That is a baseline. It costs two hours and no money, and it turns every vendor demo from an act of faith into a conversation where you already know the answer.

For the wider shift in what a marketing team can actually do now, we wrote what became possible in marketing.

Adobe placed its bet in April. Profound placed its in February. Yours fits in a spreadsheet with fifteen rows, and you can start it this afternoon.

Build your AI visibility baseline