Case Study · Leadership
A leadership story about building the operating layer before the AI mandate arrived. Not a software demo — a story about founder-mode inside a bigger org.
When you arrive to lead a growing program that’s stretched thin without central tooling — a state common to most fast-scaling operations teams — the meeting-room instinct is to observe first, then propose. That’s the manager move. The founder move is different: show up with the tool already built.
I arrived with a working scaffolding for an events operating platform — built on my own, using frontier code-generation tools. Not a prototype. A real, functioning platform I’d used at smaller scale. I enhanced it every day from day one, on my own initiative, without asking anyone’s permission to write code.
The best time to build the tool is before the meeting where somebody asks whether you have one.
A few months in, the direction was clear. As Joby’s founder JoeBen Bevirt has publicly framed it, AI is transforming “everything from aerodynamics research to back-office tasks” — a “force multiplier” across engineering, HR, finance, certification, and manufacturing. That’s a common story right now: most companies react to a top-down AI priority by scrambling to evaluate SaaS options, book vendor demos, and put together recommendations.
The events function I ran was already there. Building, not shopping. No scramble to evaluate five different AI-native tools — a platform already running, learning from real workflows, and shaping itself around the actual constraints of a global program. The moment became an accelerator instead of a distraction.
Two things I hold about AI in operating systems, both hard-earned:
One. The AI belongs in the plumbing, not the interface. In a well-built internal tool, nobody thinks “the chatbot is helping me today” — they just notice that the brief is complete, the dashboard caught the conflict early, the roster surfaced the right people. That’s the win.
Two. This kind of platform is code, not a tool subscription. Building it means the organization owns the artifact, the workflows are first-class citizens rather than someone else’s roadmap, and the AI capabilities compound over time as the platform learns from actual use. Adopting a SaaS product means renting somebody else’s idea of your operations.
Two operator instincts show up in every engagement I’ve run since:
Build before the ask. The organizations that got the most from this generation of AI weren’t the ones that bought the most tools. They were the ones that had already built the plumbing when the mandate arrived. Being ahead of the request changes the conversation from “evaluate options” to “here’s what we’re already learning.”
Frontier tools change what one person can ship. The events operating layer that would have required a small engineering team five years ago is now something an operator can build and maintain themselves — if they know how to build. That capability is a real edge for anyone leading a program in 2026 and beyond.