What it took to get off the whiteboard
I’m the General Manager of Adventure Idaho. I used to spend hours into the night working out the next day’s rafting trips. Then I’d sit down and type a long dispatch message: every guide, driver, boat and vehicle, where they were going, and when.
A reply would come back. Someone wasn’t available that day. Or someone would catch an overlap between a day-trip assignment and a multi-day trip.
So I’d go back to the board, work through what else had to move, and rewrite the dispatch. The crew was helping catch conflicts, but each correction meant another round of planning and typing.
We run day trips, multi-day trips and rentals across two bases. Our seasonal crew reached 44, and our peak day had 272 people on the water. Coordinating that much work on a whiteboard took 20 to 40 hours of planning a week.
The whiteboard was faster until we had to check it
We had booking software. It handled reservations and payments. But we were using less of it to coordinate the operation because keeping it current took too long. We needed to change the plan, not spend the evening working through screens and settings.
So we moved more of the working plan onto whiteboards and paper. Erasing a name and writing another one was quick.
After enough changes, the board was crowded with trips, names, times and vehicle assignments. The things connecting them were still in our heads. Moving a guide meant checking qualifications and other commitments. Moving a van meant checking the driver, the seats, the route and the next pickup.
The plan changed all day. A guest count changed, someone needed time off, a pickup moved. By evening, I was piecing it together again. Then I had to turn the board into a message everyone could use.
I had to write down how the operation worked
I knew the operation well enough to make those decisions. What I needed was to put that knowledge somewhere it could be checked and used without me repeating the work every night.
That became 44 interconnected tables and roughly 200,000 data points. The database holds trips, guides, qualifications, boats, vehicles, routes and assignments. The operating model connects them with the rules and calculations that govern our day.
A van has a seat count, a driver needs the right qualifications, and a trip has a minimum number of vehicles needed to move its guests and crew. A driver already heading toward a takeout may be able to cover another movement, if the route and timing work.
Boat capacity depends on the boat and the kind of trip. A multi-day trip holds people and equipment for its full length. Route miles, distances between points, travel times and time to unload and reset all affect what can happen next.
Here’s a simplified example: a group adds guests. That might require another boat, another qualified guide and more vehicle seats. Before using them, we need to know when they return from their previous trip and whether they can reach this departure in time. Those are connected checks.
I put the rules into tables that the software could read. Scripts do the calculations and check for conflicts. API connections bring booking records into the system, and syncs reconcile changes so we aren’t maintaining the whole plan by hand. Checks run when records change and again on a schedule.
That was the substantial work: turning what I knew about running this business into relationships, rules and math that could be used consistently. When someone corrects a result, I still have to establish whether we’ve found a missing rule, an exception or bad information.
Then the whiteboard became easy to use on a screen
Once that foundation existed, I could put a simple board on top. I can click and drag assignments, see conflicts, look at other options and resolve them. I can ask for a proposed crew and rig, then confirm it. The system checks qualifications and availability while I work.
The crew got web apps they can open on their phones. Guides see their trips, driving runs, training and time off together. The office and the crew are looking at the same plan. I don’t have to type the board into a dispatch message for everyone to know where they’re going.
Assignments and trip logs also leave a record we can reconcile. Who guided which trip and who drove which movement are tied to the work, rather than reconstructed later from messages and memory.
Our team chats connect to the same records. AI can read a scheduling change in a message, work out which assignments it concerns, and update an internal record or stage a change for review. The conflict checks then show what needs attention. I can also ask questions or describe a change in ordinary language.
Anything touching money, compliance or safety requires human confirmation. The system can carry routine information and do repeated checks while we make the decisions that need judgment.
Planning takes hours instead of evenings
Weekly planning is now 2 to 4 hours. The long dispatch write-up is gone. The crew checks the board, and the safety briefing still happens.
I spend more time putting the right guides with the right guests and building crews that work well together. A change still needs thought, but it no longer means rebuilding and retyping the whole plan.
You can see the board and crew apps, and how the database, connections and conflict checks work.
Where this goes next
I’m now helping business owners get their operating rules out of their heads, record them, and design connected databases that make the complexity manageable. They know their businesses. I help turn that knowledge into something the team and its tools can use.
Once our records and rules were in place, AI became useful quickly. It had current information to work with and boundaries for acting on it. That connection is what I’m describing in Everyone bought AI. Almost nobody connected it to the business.
I’ve spent weeks setting up booking software, then faced the same work again. Trip details, customer-facing and internal descriptions, pricing, group sizes and capacities, seasonal schedules, departure times and photos all have to be entered. We already track much of that information, and our systems keep the operational records current. I want AI to prepare the setup from those records instead of making us type them in again.
Booking information comes in today, but capacity, availability and prices don’t yet flow back. I want a supported way to send approved updates and read the result back to verify it, while continuing to buy checkout, payments and reservation tools.
I’m compiling what outfitters need from booking software and talking with software companies about what’s possible. If you run an outfitting business, I’d like to compare notes: when a trip changes, what do you still update by hand, and what does someone have to remember?
If coordinating your business still depends on someone remembering, retyping or chasing changes, here’s how I work with service businesses.