The marketing director at BMW has a problem that didn't exist two years ago. When a prospective buyer asks ChatGPT which luxury sedan offers the best value, will BMW even make it into the answer?
It's the kind of question that sounds almost absurd—until you consider that Google's AI Mode now fields 75 million queries daily, and Perplexity's valuation has swelled from $9 billion to a reported $14 billion in barely a year. Suddenly, the old playbook—SEO keywords, paid search campaigns, social media buys—feels inadequate. Maybe obsolete.
Enter Sitefire, a San Francisco startup emerging this week from Y Combinator's Winter 2026 batch with a proposition that's either prescient or premature, depending on whom you ask. The company promises what founders Jochen Madler and Vincent Jeltsch call "GEO on autopilot"—Generative Engine Optimization for a world where AI agents, not search engines, broker the relationship between consumer and brand.
They've already signed enterprise clients, according to Y Combinator. BMW. Xtrackers. Asset manager DWS. The logos sit on Sitefire's homepage like trophies, though details remain scarce—no case studies yet, no public testimonials. Just a LinkedIn post and the implicit endorsement of being accepted into Silicon Valley's most influential accelerator.
The visibility gap
Here's what Sitefire actually does, stripped of the jargon: it monitors whether your brand appears when someone queries ChatGPT, Gemini, Perplexity, Google AI Mode, DeepSeek, Claude, or Grok about products in your category. Most early GEO tools stop there. Sitefire, at least in theory, goes deeper.
The platform reverse-engineers each AI-generated answer, tracing it back through the web searches, cited pages, and source material that influenced the model's output. Then it gets tactical. Which third-party sites should be mentioning your brand but aren't? Which competitor blog posts are AI models citing repeatedly? Sitefire flags the gaps, recommends partnerships with editorial platforms and user-generated content communities (yes, even specific subreddits), and generates AI-optimized content designed to fill those voids.
Upload it to your CMS, the company suggests, and watch your "AI visibility" climb across 180 markets.
It's a compelling pitch. Whether it works at scale is another matter.
An inflection point, or hype cycle?
The timing is deliberate. Google CEO Sundar Pichai told investors in the company's Q3 2025 earnings call that AI Mode is "driving incremental total query growth"—corporate speak for: people are using this thing, a lot. Perplexity's rumored new funding round, reported by the Financial Times and Wall Street Journal, values the startup at nearly $14 billion, up from $9 billion just months earlier.
Meanwhile, research from Yext—a company that has its own stake in the GEO market—analyzed 6.8 million AI citations and found something startling: 86% came from brand-managed sources. Your website. Your listings. Your structured data. If you're not curating your digital footprint with AI agents in mind, you might as well not exist.
BrightEdge, which surveyed more than 750 marketers in mid-2025, reported that 68% are already tweaking their strategies to account for generative AI. And then there's Adobe's $1.9 billion bet on Semrush, announced last November and expected to close by mid-2026—a signal that the enterprise marketing stack is being rebuilt for an AI-native era.
But here's the catch: nobody really knows what GEO best practices look like. The discipline is maybe 18 months old. There are no decade-long case studies, no consensus on what metrics matter, no established ROI models. It's the Wild West, and Sitefire is selling picks and shovels.
A fragmented landscape
They're not alone. BrightEdge, seoClarity, and Yext have all bolted AI search optimization features onto their platforms. Wix rolled out an "AI Visibility Overview" for its website builder users. Azoma, a UK-based startup, raised $4 million to simulate prompts and generate GEO content for clients including Mars, Colgate, and Procter & Gamble, per Business Insider.
The problem? Most of these tools rely on sampled or modeled data rather than direct API access to the AI models themselves. Transparency is minimal. Pricing is almost never published—Sitefire's FAQ mentions "cost per prompt" and "minimum contract duration" but offers no numbers. The company's terms of service indicate subscription billing through Stripe, with monthly or annual options, but specifics remain behind a sales call.
And then there's the ethical question. A piece in The Guardian late last year warned about the "chatbot optimization game," questioning whether AI-generated answers can be trusted as brands learn to manipulate them. It's SEO's Black Hat Era all over again, except this time the stakes might be higher. Search engines surface links; AI agents deliver answers. One invites scrutiny. The other presumes authority.
The founder profile
Madler comes from Germany's top finance program—valedictorian, with Stanford coursework in deep reinforcement learning tacked on for good measure. Jeltsch is a former robotics engineer at RobCo who, according to his Y Combinator profile, once "hacked German healthcare bureaucracy." (The specifics of that hack remain unexplained, which is perhaps the point—founder mystique plays well in YC demo rooms.)
Their company structure hints at ambition beyond Europe: Munich headquarters under pulse Energy GmbH, with a Delaware holding entity layered on top. The branding is sleek, the messaging confident, the customer logos impressive.
But logos aren't case studies. And confidence isn't proof.
The unresolved questions
What Sitefire is really selling—what every GEO tool is really selling—is insurance against invisibility. As AI agents become the default interface for discovery, brands face an uncomfortable possibility: irrelevance by omission. Not because your product is inferior, but because ChatGPT didn't mention you.
Can Sitefire deliver measurable ROI? Can marketers actually tie improvements in "AI visibility" to pipeline growth and revenue? Will enterprises pay premium prices for action-oriented GEO platforms when cheaper monitoring-only alternatives flood the market?
The answers aren't clear yet. Maybe they won't be for years.
But the shift is undeniable. Google AI Mode alone is approaching the daily active user count of a major social platform. Perplexity's valuation suggests investors believe this isn't a fad. And Adobe doesn't drop $1.9 billion on a whim.
Sitefire is betting that "GEO on autopilot" will become as essential as SEO software once was. And judging by the logos on their homepage—BMW, Xtrackers, DWS—some of the world's largest companies are at least willing to hedge their bets.
Whether that hedge pays off remains an open question. For now, the marketing director at BMW probably isn't sleeping much better. But at least someone is promising to help.
