CMOs have spent the past year hearing that their brands are vanishing from ChatGPT, usually from a company selling the fix.

In a post on X this week, Harpreet Chatha argued that the companies selling answer engine optimization (AEO) software are leaning on fear, uncertainty and doubt because their revenue isn't growing fast enough. "If you know numbers, the math isn't mathing," he wrote.

He named names. Citing Ramp's spending data, he said Profound's adoption has barely grown since January and AirOps is declining. He said Clearscope had collapsed and questioned Searchable's $85 million valuation, which he set against roughly $5 million in annual recurring revenue. Peec AI was his exception, and he pointed to its new $20 million ARR milestone on about $29 million raised. He also criticized an unnamed New York startup for staying quiet about its ARR, and said several former prompt trackers have moved into content generation and ad tools. He called the post "not scientific or investment advice."

We checked his claims against Ramp's public data and the funding figures on record. The trend he describes is real. How bad it is for any one company is harder to tell from the numbers he used.

Ramp shows buyers moving to Peec

Ramp ranks AEO vendors using card and bill-pay spending from its business customers. On its AEO category page, Profound is used by about half of the businesses that pay for an AEO tool. Peec AI is second with roughly one in five. Peec's share is rising by about a percentage point a month, and Profound's by about a tenth of a point.

AirOps is third at 16%, 4 points lower than a year ago, according to Ramp's AirOps page. Clearscope is at about 10%. In March, Ramp said Profound's 40% share was about double that of AirOps and Clearscope, so Clearscope's share has fallen by roughly half since then.

That supports most of what Chatha described. Profound is holding its lead without adding much share, AirOps and Clearscope are slipping, and Peec is gaining.

Ramp measures buyers, not revenue

Ramp's adoption rate is the share of AEO buyers using each vendor. It says nothing about revenue or contract size, and it doesn't reflect how fast the category is growing. Ramp said paid adoption of AEO tools more than doubled in the year to March, but the category still reaches only about 1% of its customers. If the number of buyers keeps rising, a vendor with flat share can still be growing quickly.

Ramp's customers also skew toward tech companies, TechCrunch has noted, so large enterprise contracts may not show up. Profound has its highest adoption among enterprise customers in Ramp's own data.

The public revenue numbers are mixed. Peec said it passed $20 million in ARR, twice its level in May. Searchable raised $14 million at an $85 million valuation in May, when it reported about $2.6 million in ARR. We couldn't verify Chatha's $5 million figure. Profound raised $180 million at a $1.8 billion valuation in September, in a round led by Sequoia and Kleiner Perkins.

Startups backed by venture capital often carry valuations far above their revenue. Chatha's complaint is that some of these will be hard to justify when a buyer eventually has to pay them.

OpenAI now reports the basic numbers itself

The pressure on these companies isn't only coming from rivals. This morning OpenAI launched Businesses in ChatGPT, which shows brands OpenAI's own count of their mentions, profile views and clicks. It leaves out prompts, so prompt trackers still have work to do. But brands no longer need a vendor to tell them whether ChatGPT mentions them at all.

That's one reason vendors keep adding products. Profound now sells agents that research and write content, plus an Ads Studio for AI search platforms. Chatha called that kind of move "article spinning." The vendors describe it as going from measuring visibility to improving it. Whatever the label, a tracking dashboard by itself just got harder to sell.

How CMOs should buy AEO tools now

Marketing leaders are the audience for the fear Chatha describes, and their budgets pay for it.

Start by testing the pitch. Telling a CMO the brand is "invisible" in AI search costs a vendor nothing, and until this week it was hard to check. OpenAI's dashboard gives brands a baseline of their own. A vendor should be able to show, using the brand's data, what it found and what it plans to change.

Look past the funding announcements. A large round tells you investors like the category. Retention and customer growth tell you whether customers keep paying. Ask for both, and be wary of vendors that won't share them.

Keep contracts short. Most of these companies are less than three years old, two well-known names are losing share, and the AI platforms have started building their own measurement. A multi-year deal signed today could outlast the product it pays for.

Our take

Chatha's post is harsh, it takes personal shots at named companies, and it reads more into Ramp's share data than the data can carry. A vendor with flat share can still be growing.

On the bigger question, we think he's right. AEO has drawn a lot of venture money, revenue across the category is uneven, and too much of its marketing depends on scaring brand teams. Ramp's data shows buyers are already sorting the field.

For brands, the answer is ordinary procurement discipline. Buy AEO software the way you'd buy any measurement tool, with short terms, proof from your own data and a clear sense of what it shows that your analytics and OpenAI's new dashboard don't.