€180 booking. €39.60 gone to commission. The hotel never met the guest until arrival. The traveler paid more than they would have at the door. Every month. For 25 years. That's the trade-off online travel agencies have defended since 2000.
38% of under-35 travelers now ask an AI agent — Claude, Gemini, Perplexity, ChatGPT — for travel recommendations before opening a booking site. That share was 4% in 2024. Google traffic to hotel listings is down 26% YoY. The SEO moat is liquefying; a new distribution surface is opening. Nilo is built for it.
Inference costs fell 94% between 2023 and 2026. What took Booking.com a 500-person content ops team now runs as one generative pipeline at €0.40 per listing per week. The cost structure that made 22% commissions necessary no longer exists. We're the first lodging channel built after that inversion.
Post-pandemic, 72% of Southern European independents say their #1 business concern is OTA dependence (Hotrec 2025 survey). They're looking for an exit. Our waiting list sits at 1,380 properties — most of them inbound, no outbound sales. Supply is volunteering to migrate.
Nilo isn't a booking site with AI bolted on. It's a stack of six agents that replace the SEO team, the content team, the compliance back-office, the onboarding org, and the fraud team — the five teams that make 22% commissions necessary.
See them running on the Operating Model page, or the full walk-through on the landing page.
Not feature-level. Not "we have a chatbot too." Three structural reasons.
Booking.com's 2025 revenue is ~$23B. ~95% of it is commission. Dropping to €9/booking would erase ~$21B overnight. No public company board approves that. The commission model isn't a pricing decision, it's load-bearing architecture for 27,000 jobs. We started without it; they can't unbuild to it.
Booking.com spent 25 years winning Google. Their 1,500-person performance-marketing team and $8B+ annual ad spend are the moat — and the cost base. Agentic discovery routes around that moat entirely. Every dollar they shift to the Demand-agent surface is a dollar that stops feeding their existing moat. They're trapped.
At 40% share of bookings for a given hotel (our milestone-2 target), we are the primary channel. The guest relationship, the reviews, the repeat-booking history, the Handshake data — they all live on Nilo. A hotelier would be leaving 40% of their business and their entire guest dataset behind to churn us. That's a harder trade than leaving the 22% channel was.
Knows how to build a payments-heavy product across three Latin-language markets. Was Maria's guest at Casa Boavista in 2023. Saw the ledger.
Spent three years inside a foundation-model lab figuring out what agent-grade RAG actually needs. Designed Nilo's six-agent architecture on week one.
Spent a decade signing hotels to the commission model. Knows every lever. Switched sides in 2024. Owns Field Ops, hotelier CAC, and the PT/ES/IT rollout.
Built the PM-per-agent operating model. Hired five of the six PMs in 90 days. Still owns the flagship Agentic Demand role until we find its PM.
GC led Hopper's Series B and has the strongest pattern-match on travel unit economics plus agentic demand. In conversations since Jan — committed to lead on signed term sheet. Target close: Q3 2026.