How 5 fleet managers built their own tools without writing code
Five fleet professionals, none of them programmers, are building their own software tools using AI. They are not learning Python or JavaScript. They are describing problems in plain English and letting the AI generate the solution. One of them cut a task that used to take 80 hours down to roughly 30 minutes.
If you manage operations, field teams, or any process-heavy function and you have been waiting for a reason to take vibe coding seriously, this is it.
️ What These Operators Actually Built
Ernie Garcia, director of fleet and business systems for Gothic Landscape, started with a batch PDF printing tool and kept going. His current project pulls data from multiple spreadsheets and automakers’ order guides to generate vehicle specs. Fleet managers get a printable output. Pricing can be shown or hidden depending on who is reading it.
“It allows us to build specs and quickly update them, kind of like little Tetris pieces.” — Ernie Garcia, Gothic Landscape
Garcia also needed a total cost of ownership calculator his team could actually use. He described his users as “mechanics first and not really analysts.” His approach: use ChatGPT to write a better prompt, then feed that prompt to Copilot, which produced a multi-tab spreadsheet with TCO calculations. One AI helped him write instructions for another.
Reed Jackson, fleet analyst for Dallas County, is working on a vehicle replacement priority system. The inputs include vehicle age, mileage, maintenance history, and utilization. The output ranks which vehicles need attention. No data engineer required.
James McKinley, vice president of operations at City Rent a Truck, is combining total cost of ownership, replacement cycles, and residual values across thousands of assets. In the past, consolidating that data would have required analysts or data engineers. McKinley knows the problem deeply, and now he does not have to translate it through a programmer who may or may not get it right.
Kristin Leary, director of fleet services at Syneos Health, is building two apps. The first helps assign available vehicles based on location, tier, drivetrain, and how long a unit has been sitting in storage. The second guides drivers through photographing vehicle condition on a schedule, creating an auditable record before the vehicle comes back for sale. Both address problems her team currently handles manually through spreadsheets.
The Numbers Worth Noting
Garcia estimated that one category of task that previously took as much as 80 hours now takes him roughly 30 minutes. The article does not specify which task. That qualifier matters, but the order of magnitude is real enough to pay attention to.
How the Build Process Actually Works
Vibe coding is not a one-shot prompt. The workflow looks more like this:
- Describe the problem in plain English.
- Review what the AI produces.
- Explain what is wrong or missing.
- Add a requirement or change a calculation.
- Test the result.
- Refine the output.
- Save the successful process and reuse it.
Don Schnell, a business consultant and former Verizon fleet member, compared it to drafting and revising a document. Once the instructions and output are correct, you save the routine and run it again rather than rebuilding from scratch each time.
McKinley’s technique for better prompts: ask the AI to act as a prompt engineer and rewrite your instructions for the best possible result, then continue the conversation from there.
“What you put in is what you’re going to get out.” — James McKinley, City Rent a Truck
Where to Start If You Have Not Built Anything Yet
Garcia’s advice: start with something small and already familiar. Take data you know well, run it through an approved AI environment, and ask it to calculate or organize metrics you already understand, such as fleet age, maintenance costs, or fuel economy. The point is not the output itself. It is that you already know what the correct answer looks like, so you can catch errors immediately.
Every operator profiled here started with a manageable, manual task before expanding into bigger problems. None of them started by trying to replace their entire workflow at once.
⚠️ The One Risk Nobody Is Talking About
Schnell raised a concern worth taking seriously. The people experimenting with these tools today learned their jobs before generative AI existed. They spent years in spreadsheets, reviewing maintenance data, and building replacement analyses by hand. They know when an answer looks wrong.
“We need those users to learn how to manually look at the data and understand what good data, what’s bad data before they just solely rely on the results of AI. Failing to develop that foundation would be a fatal mistake.” — Don Schnell
Future employees may encounter the AI interface first. If they never developed the underlying domain knowledge, they will accept a confident-sounding wrong answer. The solution is not to slow down on AI adoption. It is to make sure domain knowledge and AI proficiency grow at the same time.
That is a problem every operator building with AI should think about, not just fleet managers.
I’m founder of BizStack at Cagri Sarigoz LLC and a passionate advocate for entrepreneurs.
With over 14 years in tech, marketing, and AI, including my role as Head of SEO at CitizenShipper and co-founder of TaleBot at Intale AI, I’m dedicated to sharing genuine, useful product insights and tips.
At BizStack, I aim to cut through the digital noise to provide clear, actionable advice.
And more than all else, I’m a father to a (always) little girl and a husband.
Contact me at [email protected] for assistance.
