I built CaptainBook because I watched something that shouldn't be broken: activity operators losing revenue to systems that don't understand their business.

ChatGPT can book you a flight. Claude can find you a hotel. Perplexity can plan your entire itinerary. But ask any of them to book you a sunset catamaran tour in Santorini, a diving expedition in the Red Sea, or a paragliding adventure in the Alps, and they'll fail. They'll give you generic advice. They'll direct you to Google. They'll apologize and tell you to "contact the operator directly."

Why? Because most activity operators, the thousands of tour companies, diving schools, adventure guides running bookings on legacy platforms, are invisible to AI.

The Invisible Economy

The activity booking market is massive. Over $180B annually in bookable experiences globally. But unlike flights (GDS standardization) or hotels (OTA infrastructure), there's no unified layer connecting activity operators to AI assistants.

This isn't a technology problem. It's an infrastructure problem.

Every AI model capable of making autonomous booking decisions, whether it's ChatGPT's Canvas interface, Claude's computer use, or tomorrow's native AI assistant, faces the same friction: activity inventory is fragmented, non-standardized, and unreachable at scale.

So operators miss revenue. AI assistants frustrate users. The opportunity gets stranded.

The Gateway Pattern

At CaptainBook, we've been solving this from the operator side for half a decade. We help tour operators optimize pricing, manage capacity, forecast demand, and execute revenue intelligence through AI agents.

But we realized something important: the real opportunity isn't selling operators better booking software. It's making their inventory bookable by every AI assistant that exists.

We call this the Agent Transaction Gateway. It's infrastructure that:

1. Standardizes operator inventory into a format AI assistants understand natively

2. Makes every booking actionable, pricing, capacity, demand signals, all available to AI in real-time

3. Handles the hard parts that AI can't: payment processing, operator communication, booking execution

Why This Matters Now

The timing window is narrow. Right now, in early 2026, AI assistants are racing to become the default interface for bookings. Booking.com, Google, Expedia are all adding AI co-pilots. ChatGPT is starting to make reservations. Claude has been testing booking capabilities.

The operators who make their inventory AI-accessible in the next 6-12 months will capture disproportionate share of AI-originated bookings. The ones who wait will find themselves competing for a shrinking pool of human-originated demand.

This is how distribution shifts. Not overnight. But decisively.

What We're Building

We're launching the Agent Transaction Gateway this year. It will:

• Enable any activity operator to become AI-bookable without leaving their current booking platform

• Create a network effect: more operators = more inventory = more valuable LLM integrations = more bookings

The Question for Operators

If you're running an activity business, here's what I'd ask yourself:

By next year, will your booking system be ready for AI? Or will you still be relying on Google Search, travel blogs, and manual inquiries, while AI assistants handle bookings for everyone else?

The infrastructure is coming. The question is whether you'll be visible when it arrives.