Promptwatch raises €6M Seed to optimise brands for the age of AI-generated search

Promptwatch, an Amsterdam-based AI marketing startup founded in 2025 by Gijs de Groot (CEO) and Klaas Foppen, has raised €6 million in a Seed round led by seed + speed Ventures, with participation from Blum Ventures and follow-on from Arches Capital, which backed the company's pre-seed in September 2025. The capital goes to product development, international expansion, team growth, and a new New York office.

Measuring visibility in the new search layer

Generative Engine Optimisation — GEO — is what happens when a brand asks not "where do we rank on Google?" but "do we appear in what ChatGPT, Claude, or Perplexity says about our category, and what does it say?" The answers are inconsistent, hard to audit, and change as models update. Traditional SEO tools do not measure them.

Promptwatch's platform processes more than 10 million data points per day across the major AI search and answer engines, showing organisations how they are represented in AI-generated responses — and, importantly, where they are absent or misdescribed. The product then closes the loop: content gap identification, AI-ready content generation, and automated publishing to CMS, so the fix reaches production without requiring a separate workflow.

The company counts 1,840+ organisations as customers, including Duolingo, Fireflies, and Monks. It reached €2 million in ARR within twelve months of launch — a data point that matters not because it proves scale, but because it confirms that marketing teams are actively paying to solve this problem right now, not treating GEO as a future concern.

The logic behind a €6 million Seed

"In less than a year, we've built advanced capabilities including our Content Agent, Actions and Model Context Protocol (MCP), which connects Promptwatch to AI tools such as ChatGPT and Claude," de Groot said. "With Agentic AI Search Optimisation, we're making it easier for organisations to scale their AI search marketing through automated execution."

The MCP integration is a meaningful detail. Model Context Protocol is the standard that AI systems use to pull in external data — building Promptwatch as an MCP-connected data source makes it a potential live input to the same AI engines it monitors. Whether that creates a defensible technical advantage or is simply table-stakes for the category is not clear from the round announcement.

seed + speed Ventures led the round; Blum Ventures and Arches Capital (follow-on from pre-seed) participated. The decision to open a New York office alongside the European team reflects where a significant portion of the 1,840 customer organisations are based: US enterprise marketing budgets are larger, and US digital teams have moved faster on AI tooling adoption than their European counterparts.

The 18-month test

The GEO category is real. Whether Promptwatch is the enduring category leader is the open question. Google — whose own AI Overviews are part of the problem Promptwatch monitors — will almost certainly build analytics for AI visibility into Search Console or similar tools. SEMrush, Ahrefs, and Moz already offer early GEO features. The incumbent distribution advantage is substantial.

Promptwatch's defensibility argument rests on depth and speed: 10M+ data points per day, a more complete model of AI search behaviour, and enterprise relationships already signed before the incumbents catch up. Whether that lead holds over 18 months, and whether the agentic layer it is now building creates a workflow integration too expensive for clients to switch away from, is what this round is buying time to prove.