AI Builder Stories
Behind every AI tool is a person deciding which problem is worth spending time on. Luke Grace built BYOBot around a simple question: how do you make useful automation available to people who do not already know how to build it?
What Luke built
BYOBot turns a routine described in plain English into a portable agent playbook: prerequisites, ordered steps, smoke tests, and fallbacks. Rather than a hosted integration, it is a specification that can run in the AI tool someone already uses. The first agent is free and requires no account.
Why he built it
Luke had spent years building automations for repetitive work in his own field. The work itself could be automated. What was missing was an on-ramp for someone without his technical background to reach the same result and train an AI agent to take a task off their plate.
One lesson for builders
“The distance between ‘I can automate this for myself’ and ‘anyone could automate this’ is far shorter than it looks from the inside.”
For Luke, the hard part was not automation. It was making it approachable at any level of expertise. He also learned that an agent which only completes a task is half finished. It should QA its own work, or the person using it has simply inherited the checking. That is why BYOBot builds testing into the playbook rather than adding it later.
Explore BYOBot, and browse other practical tools in The AI Navigator’s AI tools directory.