Google Built the SEO Industry. AI Is Dismantling It.

01 /The Concept

Google's answer to a question is a list. Ten links, all competing for the same click, and the brand that isn't there misses the consideration entirely but can still be found by anyone who keeps scrolling. The AI model's answer to the same question is a sentence. One brand gets mentioned; the rest don't exist in that response.

This isn't a future problem. The shift is underway now, and Promptwatch, founded in Amsterdam in 2024, is the platform built to manage it. The product tracks brand visibility across ChatGPT, Claude, Perplexity, and other AI answer engines; identifies the sources and content gaps that determine how AI models perceive a brand; generates AI-optimized content to close those gaps; and publishes improvements automatically through CMS integrations for WordPress, Webflow, and Framer.

The platform processes more than 100 million data points daily. By May 2026 – twelve months after launch – it had reached €2 million in annual recurring revenue and nearly 2,000 users, including Duolingo and Fireflies. The team of 17 raised a €6 million seed round in September 2026, led by Berlin's Seed + Speed Ventures, with Blum Ventures and Arches Capital participating.

The company's main named competitor is Peec. The meaningful difference: Peec tracks AI brand visibility well; Promptwatch also acts on what it finds, generating and publishing content to CMS platforms automatically. The combination of tracking and automated publishing is what Promptwatch considers its durable advantage – each feature exists in other tools; having both, integrated, is harder to replicate than either alone.

02 /Why It Matters

The standard framing of this space – 'SEO for AI' – implies the same fundamental job, adapted for a new channel. Promptwatch's product design reveals why that framing is wrong.

Traditional SEO operates at the metadata layer: titles, keywords, anchor text, structured data. The mechanism is well-understood – search engines crawl pages, index signals, and rank based on a reasonably transparent set of factors. AI model perception works differently. Language models form impressions of brands from their training data and retrieval context, weighted by recency, citation frequency, and the perceived authority of sources that mention them. Gaming a keyword is relatively tractable; shaping what a language model 'knows' about a brand requires ensuring the right information is in the right places in formats models can consume and cite.

The 100 million daily data points Promptwatch processes aren't website crawls. They're AI model queries – asking AI systems what they believe about brands, tracking how those beliefs change, and identifying which content changes moved the needle. That's measurement of a genuinely new signal, not an adaptation of a metric that already existed.

The most consequential design decision is the closed loop. Knowing your brand is absent from AI answers to a relevant query is inactionable if acting on it requires a six-week content production cycle. The CMS integrations that turn an insight into a published fix in hours rather than weeks are the product decision that distinguishes a real workflow tool from a visibility dashboard.

03 /Opportunities

The immediate expansion path is enterprise, where the business case is clear but the go-to-market is still forming. Promptwatch's current ~2,000 users at roughly €1,000 average ACV have validated the product at the growth-stage company tier. Large brands with multi-market presence, multiple product lines, and existing SEO functions already paying six figures for tools and services are a natural upgrade path – same workflow, 50–100x the contract size, with managed services layered on top.

The agency version is the more structurally interesting play. The global SEO agency market – roughly $25 billion – is in the middle of reorienting toward AI search optimization without a clear toolset for operating it at scale. A white-label or agency-seat version of Promptwatch that lets a digital marketing agency manage AI visibility for 50 brands simultaneously, with standardized reporting and account-level analytics, is a different distribution model than selling directly to brands. Agencies have the client relationships, recurring billing, and incentive to add services; they just need the tool.

The most specific near-term position worth taking: AI visibility optimization for high-consideration B2B purchase categories – enterprise software, fintech, insurance, healthcare – where being absent from an AI model's answer in a buying process means a missed deal. These buyers are actively using AI systems in their purchase research today, which makes the urgency real and the ROI directly calculable. That specificity produces a better sales conversation than a general AI brand visibility pitch.