I was wrong about AI. Dead wrong. As AI gained momentum and companies (like mine) began reshaping their product roadmaps around AI capabilities, I dismissed it as another passing trend with limited practical applications. I’d seen this pattern before—technologies surrounded by hype cycles that promised transformation but delivered unclear ROI, often increasing costs while creating new expenses like sourcing scarce, expensive talent. At the time, our public company status heavily influenced my perception of our AI investments, and I believed we were primarily responding to market expectations rather than actual strategic value. The pressure to maintain our stock price seemed to drive our AI initiatives more than any clear vision of how the technology could meaningfully benefit our business.

Fast forward to today, and AI has become integral to my daily workflow. I rarely go a day without opening Claude Code or consulting our enterprise AI solution. The turning point came when I first tried Claude Code. While I had a ChatGPT subscription, the ability to interact with AI directly from the command line—pointing to files in my filesystem and working within my native development environment—was revelatory. Instead of constantly switching contexts and copying code back and forth, I could bring the power of AI to where I already work. I initially started using Claude Code because I believe you can’t effectively protect or govern what you don’t understand or haven’t experienced firsthand. What I didn’t expect is to rediscover my passion for building.

Ideas for applications have been flowing—some small quality-of-life improvements, others more ambitious projects. The simpler ones include automating address updates across systems after our many moves, or having files automatically named and sorted after download. But I’m also drawn to bigger challenges, like developing an open source IGA solution in Golang to tackle persistent identity management problems that today demand expensive, operationally heavy platforms.

The common thread? AI as an incredible accelerator. It’s expanded my capabilities in every direction—helping me pick up new languages, debug with speed, and venture beyond my go-to tech stack (even if C# will always hold a special place). I’m shipping secure, functional code faster than I ever imagined possible. Last week I built a greenfield microservice containerized application in Go with UI and observability stack (Grafana, Prometheus) on the backend. I was able to get that up and running in a matter of hours, instead of days. The implications for our industry are profound, and I’ll be unpacking those thoughts in the coming weeks.