“
How could I better make a schedule in a day, instead of over 4 weeks?”
Ann Findley, Transit Planner, Summit Stage
Agency snapshot
Summit Stage, Summit County, Colorado.
22 bus routes, 22 hours a day, 80 drivers.
Planner: Ann Findley
The problem
Every driver at Summit Stage works from a "paddle:" which bus, which routes, the breaks, the start times. Creating paddles while keeping drivers happy is the backbone of the operation, and since Summit Stage’s founding about thirty years ago, it has been done by hand. But what happens when that hand retires? Ann isn't looking to replace her expert; she wants to capture the craft to make the whole agency more resilient and efficient.
routes
hours a day
drivers
scheduler, by hand
“The one person that does this at our office is very good at it. She does it all in her head with paper and pencil before transcribing to Excel.”
“We go back and forth through three or four iterations before we dial into a really good schedule that the bus drivers will appreciate.”
“I looked into AI but I couldn’t really find a way to use it.”
Two ways out: six figures of software, or hope she never retires.
Ann found a third way.
Her husband, a bus driver, had encouraged her to try AI for years —
but where to start? Turns out, she just needed the right problem to solve.
What she did
1
She started from the destination
Not “automate my schedule,” but a clear end state: a paddle in a day instead of weeks. She knew what done looked like and worked back from there.
2
She gave it the department’s real work
She uploaded two past run cuts — her scheduler’s craft on paper — and let the tool ask her questions instead of hunting for a perfect prompt. “It came up with 90% of those abbreviations all by itself.”
3
She iterated the way the team already works
The same back-and-forth the department runs on every schedule, now with Claude in the loop: this works, this does not, here is the edge case you missed. Draft after draft. Her rule: “The more info I give, the more accurate it’s gonna be.”
4
She kept it in Excel
The output lands in the tool the team already uses, exported and owned in-house.
Before AI Adoption:
4 weeks
AFTER AI ADOPTION:
2 hours
to the first AI-built paddle draft.
— Ann Findley, Planner, Summit Stage
What this means for you
"All you need to do is open up AI and just talk to it or type to it. That's all you need to do!"
— Ann Findley (not a techy, but that won't stop her!)
The future of work here is not software replacing an expert. The future is AI and experts working together: the tool drafts fast, the experts make it right — years of craft captured in a tool the whole team owns, with the expert at the center.
Does your operation run on know-how that lives in one person's head? Name where you need to end up, hand AI real examples of the work, and run your normal review loop with it.
Whose head does your operation live in?
Capture the craft while the expert is still in the building.
More stories
Transit
Federal Reporting: Automating NTD
The federal report every transit agency dreads, done in 25 minutes.
Case study
Read the story
⟶
Government
Regional Strategy: Streamlining CEDS
A five-county strategy that takes months, drafted live in one afternoon.
Case study
Read the story
⟶
Transit
Service Reliability: Predicting No-Shows
Taught a spreadsheet to see no-shows coming — six calls a night, seats kept.
Case study
Read the story
⟶