Most software is built by people sitting behind desks.
There is nothing wrong with that, but it creates an interesting challenge when the users spend their day on rooftops, in mechanical rooms, crawl spaces, maintenance shops, and apartment communities.
As a Maintenance Supervisor, I learned that field technicians interact with technology differently than office workers.
Speed matters.
Simplicity matters.
Visibility matters.
A technician standing beside a broken HVAC unit doesn’t want to click through five menus to find an answer.
They want information immediately.
When I started building Fix-It Fast AI, one of my biggest goals was reducing friction.
Every extra step creates an opportunity for a user to give up.
That influenced several design decisions.
Instead of requiring users to manually enter equipment information, I focused on photo-based identification.
Instead of forcing users to search through large knowledge bases, I focused on symptom-driven troubleshooting.
Instead of presenting large blocks of technical documentation, I focused on practical next steps.
Another challenge was dealing with imperfect information.
In software development, we often assume users will provide clean data.
In the real world, equipment labels are faded.
Photos are blurry.
Model numbers are partially missing.
Descriptions are incomplete.
Building software that works under those conditions requires a different mindset.
The system has to be flexible enough to work with uncertainty while still producing useful results.
One lesson I’ve learned throughout this project is that domain knowledge matters.
Understanding how maintenance technicians think and work has been just as important as understanding AI models and software architecture.
Technology is only valuable when it solves real problems.
The best software isn’t built around features.
It’s built around the people who use it.
Top comments (0)