Insights
Short, sourced pieces on the technology decisions mid-market operators are actually facing. Every claim is cited, and where a widely repeated statistic doesn't survive scrutiny, I say so.
Google tested 117 objective metrics against what its own engineers reported as technical debt. None of them predicted it — the models explained less than 1% of the variance. What to measure instead, and why AI-assisted development is accelerating accumulation right now.
Read the article →Independent data shows AI is being paid for out of existing IT budgets, not net-new money. That turns portfolio rationalization from deferrable hygiene into the thing that funds your AI roadmap — and changes what the program is for.
Read the article →Agent adoption among large enterprises roughly doubled this year. The share of organizations reporting EBIT impact did not move. A practical test for which of your processes are genuine agent candidates.
Read the article →Deploying AI onto a weak data foundation does not produce a smaller benefit — it can produce a negative one. What data readiness concretely means, why the famous "95% of AI projects fail" number is misleading, and how to scope the work so it finishes.
Read the article →The statistic your board keeps forwarding counts a company that moved one database the same as a company that left entirely. Underneath the bad number is a real shift in how modernization gets done.
Read the article →Topics
Portfolio decisions over platform decisions. What to migrate, what to expose, what to retire, and what to deliberately leave alone.
The foundation that determines whether an AI investment has a positive or negative sign. Representativeness, ownership, definitions, lineage, freshness, access.
Which processes are genuine agent candidates, what governance actually enables scale, and how to tell a real vendor from agent washing.
Why debt resists measurement, what to track instead, and how portfolio rationalization became the funding mechanism for everything else.
The Production AI Blueprint applies this thinking to your systems and produces a sequenced roadmap — with the data, governance and controls behind it — that you own outright.