Perspectives & Insights
Thinking on technology leadership, AI, mainframe modernization, PE-backed transformation, and the ideas that shape how I work.
Somewhere in the convenience of AI, we’ve lost a step that used to be non-negotiable: validation. A well-written answer feels correct, so we stop checking. The tools have changed. The need to know what’s true hasn’t.
The boardroom question has changed. It used to be “What’s our AI strategy?” Now it’s “What did we get for it?” Most organizations don’t have a good answer — not because the data doesn’t exist, but because nobody built the model to capture it.
Every system accrues it. Every team carries it. The only question is whether you’re paying it down deliberately — or letting it compound into something far more expensive than the shortcut that created it.
85% of enterprises increased AI investment last year. 95% of pilots delivered no measurable P&L impact. The discipline that's missing is feature-level unit economics — and the CTO should own it.
A company tested the premise that AI has lowered the barrier to building software so much that you don't need engineers anymore. This is what it actually cost them — and what every executive needs to understand before running the same experiment.
If you don't have measurable ROI from AI today, you are behind. Not conceptually — competitively. A ground-level view of where AI actually works, where it fails, and how to close the gap.
What 20+ years of hiring software engineers — from fresh college graduates to industry veterans — taught me about the team that actually wins.
AI is no longer optional — any organization that doesn't develop real competence with it will be outcompeted. But adopting AI without a clear business objective isn't innovation. It's theater. A look at what responsible adoption actually means in practice, and the five disciplines that separate the organizations getting it right from the ones learning expensive lessons.
More frequently than expected, someone declares the mainframe dead. They've been wrong every time — and with enterprise AI reshaping how software gets built and operated, the conversation is shifting again. Not toward replacing the mainframe, but toward something far more interesting: making it smarter.
The most dangerous phrase in business is: "We've always done it this way." Real transformation begins the moment you're willing to question everything—including the assumptions baked into how your technology organization is structured, what it builds, and why.