I asked ChatGPT to fix my schedule. It couldn’t. Here’s what had to change.
Peak season was short staffed, and I was up late planning the next day’s rafting trips. The whiteboard held the trips, guides, drivers and shuttles. A change in one place could break three others.
I photographed the board and asked ChatGPT to help make the schedule. It suggested guides who were days away in the heart of the Idaho wilderness for half-day trips.
Useful. If we’d also invented teleportation.
I put the tool aside and went back to the board and the late nights. It was an instructive failure: the photo showed names and boxes, but it did not tell the model who was qualified for which river, who was already committed, how long a guide would be away, which vehicle could make a pickup, or what would happen if a trip changed.
The following spring brought heavier bookings. I came back to the problem determined to solve it, but the answer was not a better prompt. The business needed records for trips, people, vehicles and assignments, with relationships and rules the team could check. Only then could software help us spot a conflict and show a workable change to a person who knew the operation.
Before
The plan in marker and memory.
After
A board backed by connected operating records.
The current system flags conflicts for a manager to review. It does not make safety or staffing decisions on its own. The change came from putting the operation’s knowledge where the team and the tools could use it. That connected record of people, commitments and operating rules is what I mean by a “company brain,” not an AI that somehow knows everything.
If you are new to AI, choose one task and ask: What information does it need? Which rules must it follow? Who checks the result? A reminder may only need a rule; a messy message may need interpretation; a staffing change needs a person. That is a more useful starting point than asking a tool to understand a business it has never seen.
Read the full operating story for the work behind the board, or how I built and handed it off for the implementation. If you want to start with a task in your own business, see the practical AI path.