TubeBuddy / BENlabs · 2023 – Present
Driving AI Adoption Through Results
BENlabs was already a Utah AI leader, with predictive models embedded in the business. But its first coding-assistant trials came before the tools were ready. I closed the gap with an independent React Native proof, built in four weeks using the production codebase as a reference for functionality and APIs.
30-second read
What matters in this case
- The challenge
- A Utah AI leader had tested coding assistants before they were ready, while a critical mobile rewrite remained unfinished eighteen months into a six-month plan.
- My contribution
- I proposed an AI hackathon, then used the tools part-time to build an independent React Native proof, using the production codebase as a reference for functionality and APIs.
- The result
- The demonstration shifted organizational behavior and informed the AI-assisted foundations now supporting the web portal.
- Pivotal decision
- I chose production-shaped proof over another AI presentation, while keeping the work narrow enough to complete alongside my leadership role.
- 4 weeks
- Working React Native proof
- Culture shift
- Organization adopted new AI-assisted practices
The Situation
BENlabs was not new to AI. It was one of Utah’s AI leaders, with predictive models already embedded in the business. Generative AI for software development was a different story: the team’s first experiments arrived before the tools were capable enough for deep adoption.
That timing created a credibility gap. When the models improved, the early results still shaped expectations. Meanwhile, contractors had been porting our mobile app from Xamarin to MAUI for over eighteen months — a full year past their six-month deadline.
Early tools, lasting assumptions
The first coding assistants had not delivered enough value, and that experience continued to shape expectations after the technology improved.
Stalled contractor work
External contractors were scoped to finish the port in six months. Eighteen months in, there was no end in sight.
A real inflection point
The tooling had crossed a threshold. The question was no longer whether generative AI could help, but how to make that change visible and credible.
Building the Skill
As the tools matured, I invested deeply in understanding how to work with them. Not just prompting — building a practice.
Research synthesis
Ingesting research data, finding patterns, generating insights. What used to take days of analysis could happen in a conversation.
Rapid prototyping
Building working prototypes during user interviews — testing ideas with short feedback loops instead of waiting for engineering sprints.
Production code
Not just mockups — real React components, interaction patterns, engineer-ready handoffs. UX copy, research plans, feasibility testing.
Workflow automation
AI-powered code reviews, automated documentation, testing infrastructure. Systems that help the whole team, not just me.
The early caution was earned: the first tools made mistakes and were easy to dismiss. I kept testing each generation, learning which model fit which job and how to structure problems around its strengths. The technology improved; the skill compounded with it.
Changing the Game
The company did not need another AI pitch. It needed evidence that the tooling had crossed the line from interesting to useful. I partnered with the CTO and proposed something simple: "What if we just did a hackathon next week?"
The Proposal
The CTO was enthusiastic about AI's potential. I proposed something low-risk: "What if we just did a hackathon for a couple days?" As part of leadership, I had the standing to make it happen.
We carved out a few days. My goal was not to replace the production app. I used its codebase to understand the existing functionality and APIs, then built the React Native implementation independently to make the new capability visible and credible.
Days 1-3: Hackathon
A decent number of features were already working. Core navigation, key screens, state management — not just scaffolding, but functional pieces of the app.
Weeks 2-4: Part-time continuation
Kept building in spare hours. Features, polish, edge cases. Still doing my actual job as Head of Design.
Week 4: The reveal
Shared a nearly complete working proof with the broader leadership team. Some aspects were better than the original. Their reaction: "Holy crap."
What Changed
The mobile proof got attention. But what mattered more was what it unlocked — permission to build things differently.
Labs became the future
We started building a new web portal as a "labs" area — a place to test ideas and prototypes. But the team saw what was possible and wanted to move to it immediately. It became the main environment for building new production features for customers. See how we turned that workflow into infrastructure →
Production-grade foundation
Working with Nick (who handled most of the CI/CD), we built a solid foundation: modern frontend stack, light/dark mode, design system with Storybook, analytics, localization, frontend linting. AI assistance accelerated everything.
Team adoption
AI code reviews. AI agents for routine tasks. Claude Skills for specialized workflows. Context documents for project understanding. The whole team started working differently.
What It Demonstrated
The CTO later wrote an internal proposal for broader AI adoption across the organization, using my work as an example of what's possible when someone invests in building the skill. He called it "The Britton Effect" — what happens when someone invests in AI as a real skill, not a toy.
It's not that I'm uniquely talented at coding. It's that I understood the tool deeply enough to apply my actual expertise — design thinking, product sense, user understanding — at a pace that wasn't possible before.
AI doesn't replace expertise — it amplifies it. But only if you build the skill.
What I Learned
Prove the inflection point
Early disappointment can outlast the technology that caused it. A result people can inspect makes new capability concrete.
Build with believers
The CTO already understood AI’s potential. Partnering with someone who saw the opportunity created room to turn curiosity into organizational proof.
The skill compounds
Early investment in understanding AI deeply pays dividends. Each project teaches you something that makes the next one faster.
Technology choices matter
The contractors chose Xamarin → MAUI. I chose React Native. Picking the right foundation matters more than how hard you work on the wrong one.