Product2 publishers3 min readPublished
Profound raises $180M to sell brands a sampled view of what buyers ask ChatGPT
The Series D values the two-year-old marketing startup at $1.8 billion. Its prompt data comes from opted-in consumer panels, and the mention-frequency score a brand sees on the dashboard is a sample estimate.
The Product Desk · Product desk
What happened
- Profound said Tuesday it has raised a $180 million Series D at a $1.8 billion valuation, jointly led by Sequoia Capital and Kleiner Perkins with Lightspeed, Khosla, Saga Ventures, Evantic and South Park Commons joining.
- The round lands less than seven months after a $96 million Series C, at a company that launched two years ago.
- The platform identifies the prompts a brand's target buyers use to research products, then identifies ways to raise how often the brand is mentioned in chatbot answers.
- Profound says revenue has tripled in the past six months and that it now has more than 1,000 enterprise customers, among them Comcast, The Estee Lauder Companies and Walmart.
- The company will spend the money post-training AI models for marketing use cases and building a benchmark for how well those models automate the work.
Compiled by The Product DeskSomething wrong?How this is made
Why it matters
- constraint Measurement depends on panelists who happen to research your category, so a narrow B2B buyer set may generate too few shared prompts for the dashboard to say anything reliable.
- capability Competitive monitoring becomes a weekly list of the specific prompts where a rival gets named in the answer and you do not.
- cost Justifying this price means far more spend per account than today, and the enterprises already signed are the ones who will be asked for it at renewal.
- precedent If buyers adopt Profound's planned benchmark, the standard for judging automated marketing work will come from a company selling the automation.
A category manager at a consumer electronics brand opens the dashboard on Monday and sees last week's most popular smartwatch prompts, with a note that many of them ask about health tracking features [6]. That screen is the product. The prompts come from consumer panels: people who agree to share their AI queries with market researchers through what Profound describes as a double opt-in, with a filter that strips prompts containing personal data [5].
So the mention-frequency number is a sample of what panelists asked. Nobody is counting what a brand's own buyers asked. It matters on Friday, when someone senior asks why the number moved and the brand has no log of its own to check it against.
Align site content with those prompts, and according to SiliconANGLE, chatbots cite that content more often [7]. Neither published account says how much any named customer's mention rate rose [3].
The feature I would buy first is the dull one. Profound also checks whether AI answers describe a company's products accurately, and it scores consumer sentiment around each offering [8]. Wrong compatibility, a discontinued item still being recommended: that is a defect list, and a support lead can work it this week without adopting any theory of AI attribution.
TechCrunch places the company in a wave of GEO and AEO startups trying to surface products inside the AI systems people increasingly use for search [17]. On price, the new round is about 1.9 times the size of the Series C [1]. Divide $1.8 billion by the enterprise customer count and the price is under $1.8 million of valuation per account [2]. Reaching it needs those accounts to spend a lot more than they do now.
The money is going into agents. Profound debuted a feature called Aim earlier this year that plans AI visibility work on its own, then shipped a tool that generates chatbot ads, and in June added one that measures how a brand's chatbot ads stack up against competitors' campaigns [10][9]. "Profound started as an analytics platform for marketers to understand how buyers discover their brands through AI," co-founder and chief executive James Cadwallader said [11]. "Today, they're using our Agents to research, write and report" [12].
Two questions sort this for most teams. Does your own analytics already show buyers arriving from AI answers, and do the pages the chatbots cite belong to you. Both yes: the seat pays for itself inside content work you were already doing. AI referrals but citations sitting on review sites and retailer listings: the report names a problem your CMS cannot fix, and the budget belongs to PR and retailer relations. No visible AI referrals but pages you control: take the cheapest monitoring tier and look again in two quarters. Neither: you are buying a market research report.
What to watch
- Whether anyone who is not selling AEO software adopts the benchmark Profound plans to build for automated marketing work.
- First-party citation reporting from the chatbot vendors themselves, which would compete directly with panel-sampled prompt data.
- Renewal behaviour at the 1,000-plus enterprise accounts once finance compares the seat cost against measured AI referral traffic.