Booking software solved the transaction. Nobody solved the operation.
Every guided-experience business runs on booking software: Flybook, Peek, Xola, Checkfront, FareHarbor, one of a dozen others. They are good at what they were built for. A guest finds you, picks a date, pays, and gets a confirmation. That problem is solved.
Then the booking lands, and the actual operation starts. Who guides it. Who drives. Which vehicle, which trailer, which put-in. Is that guide licensed for that section at that flow. Is the van already committed two canyons away. Did the shuttle leave in time. Did the balance ever get charged. None of that is in the booking software, and in most companies it lives exactly where it lived thirty years ago: on a whiteboard, in a group text, and in one person’s head.
I know this because I ran a multi-base, 44-person outfitter that way, and then I stopped.
The AI that shipped went to the wrong side of the business
Here is the thing I keep noticing. Every AI feature I have watched ship in this category is guest-facing. Chat that takes a booking. Upsell engines. Automated review requests. Dynamic pricing. All of it points at the customer, because that side has a clean, uniform story: spend money here, get more bookings, and it works roughly the same for every operator.
The operation got almost nothing. The products that do reach into operations have been far slower to ship anything comparable. And that is not an oversight, it is economics. The demand side is uniform and the ROI is legible. The operations side is idiosyncratic: every operator has different trip types, different rigs, different crew rules, different permits, different states. High integration cost, messy per-customer logic, small market. A product company looks at that and walks away. An operator looks at it and can’t build it alone.
So it sits there, unbuilt, which is exactly the kind of gap that doesn’t get closed by a feature request. It gets closed by someone who has lived inside the operation.
What it looks like when you build it
I built the operations layer for the company I run, on top of the booking system’s API. The booking data flows in. Everything downstream runs on a system the whole crew can see.
The part that matters most, and the part booking software has never touched, is conflicts. Every operation’s rules are different, so the system lets you define your own: which assignments collide, which qualifications a trip requires, how capacity is capped. It catches a double-booked van or an unqualified guide before the boat ramp, not after. And increasingly I resolve those assignments by voice, talking to the system from the truck instead of hand-editing records. The days of manually updating every field every time something changes are the days I was trying to leave behind.
The line I hold everywhere: a human stays in the loop on anything touching money, compliance, or safety. The AI does the coordination and the bookkeeping. It never gets to gatekeep in front of a safety call. That constraint is not a limitation. It is the reason a crew will actually trust it.
Why this is the seam
Booking software is a solved, mature, well-funded category. The operations layer underneath it is none of those things, and the same shape of problem exists in every operationally complex service business that runs on bookings: field services, hospitality, marinas, events, tours. The company that figures out how to let an AI operate safely on a shared, structured database of a real operation is building the thing the last decade of SaaS left on the table.
I don’t think that gets built from a slide, or by a team that has never had to get a van to a takeout with no cell service. I think it gets built by people who have run the operation and know exactly where it breaks. That is the work I’m interested in.