Your business shouldn’t depend on you remembering everything.
I’m Justin Smith, an operator and systems builder. I map how a business really works into connected data its teams and AI can actually use. At Adventure Idaho, where I’m General Manager, I lead the operation and built the system it runs on.
- 3.6×customer growth
- Nearly 20×website traffic
- 20–40 → 2–4 hrsweekly planning, estimated
Sources and estimates
- Customers: 1,279 in 2022 → 4,709+ in the 2026 season so far.
- Web analytics: 4,944 → 97,486 annual unique visitors; 2022 vs. Oct. 10, 2025–Oct. 9, 2026.
- Weekly planning: estimated 20–40 hours → 2–4 hours.
- Revenue and other financial figures are kept private at the company’s request.
- Crew + trips Colored assignments put each guide on a trip or mark time off.
- Driving Shuttle entries carry the vehicle and pickup point.
- Conflict checks The red outline flags a conflict for a person to review.
One change. Everything it touches.
An example: a group asks for a later pickup. Follow the request through the operating system.
- A request comes inA crew message or form captures the proposed pickup change.
- Connect the recordsMatch the trip to its guests, guides, vehicle and shuttle. Booking data comes in read-only.
- Check the rulesTiming: does the next trip still fit? Capacity: are there enough seats? Qualifications: can this guide and driver do the work?
- A person reviewsSurface conflicts and verify the proposed change. Human approval for money, compliance and safety.
- Update the operationThe approved pickup time updates connected operating records, the board and crew tools. Report back what changed.
An operating change should only need to be made once. I’m working with booking providers on the connections needed to make that possible. Our current integration reads booking data; approved write-back is the next step. What booking software needs to support.
The hard part is deciding what the system should know, what it should do, and where a person stays involved.
The order matters
- The product. Make it the same every time, delivered by people who are not you.
- The people. Training, standards and debriefs, so quality holds while the volume grows.
- Demand. Get found, get trusted, get chosen.
- Pricing. Raise it once demand has proved the product is worth it.
- The operating system. The process and the software underneath it, so growth stops running through one person's memory.
Most of the damage I see comes from doing these out of order. A demand engine pointed at an inconsistent product just manufactures bad reviews.
What I built
- Half the revenue base gone in FebruaryA policy change took the largest account. Records by May.
- The architectureThe operating platformOne database the whole crew runs from. Assignments made from a phone, updates in a sentence, and every change checked against everything else.
- The demand engineFrom invisible to highest rated in the market. Content and reputation built the foundation; paid marketing now helps scale demand.
- The sensor layerBookings, forms, radios and crew messages, read and reconciled against the live operation.
- Building a training program from nothingThe standards and debrief culture that held quality steady while the company tripled.
What I have figured out
- Start here: the whiteboard storyWhat it took to get off the whiteboardLate nights at the whiteboard, long dispatch messages, and the database that made our rafting operation easier to run.
- Build it centralized. Run it distributed.Why the system that replaced our whiteboard got built in weeks, and the three phases I'd use inside any service business.
- Everything is connected except the booking softwareWe run on our own system with AI on top. What we need from booking software, and why it's bigger than one outfitter.
- I was the APIFor a few seasons most questions in the operation routed through me. What it cost, and how it stopped.
- Selling through machines to peopleMachines now decide who gets considered. What they read, and soon act on, is the operation underneath.
- From about 30 to about 1,800 AI search appearances a monthThe lever wasn't more content. It was an operation worth describing.
- The expensive part was never the softwareWhy custom operating software never reached small businesses, and what changed.
- We stopped hiring for experience and started hiring for trajectorySeasonal crews in a town nobody commutes to. What we changed and what it cost to learn.
Right now
- The season. Closing out a record season at Adventure Idaho, and building the systems and training to run the next one better.
- The research. Working out where a business's unwritten rules actually live, so an install starts from evidence instead of interviews.
