Writing
On data & analytics
Building and leading data & analytics — leadership, platforms, machine learning, and the frontier: agentic analytics and LLM evaluation.
-
Nobody Asked for a Chatbot
Teams are building their own AI analysts, and the data modeling underneath is still the hard part. But in my experience, adoption depends on how findings reach people. Most users want answers waiting for them, power users want analyst-grade access, and almost nobody has asked me for a chat box.
-
The Token Bill Is an Architecture Problem
At July's costs for our question enrichment pipeline, aligning our question bank to multiple state standards frameworks was projected at tens of thousands of dollars. The full run cost a few hundred, and quality held. None of the savings came from cheaper token prices alone. They came from the same moves that fix a warehouse bill.
-
The Work That Makes AI Coding Agents Pay Off
A study of 700+ software companies found AI coding agents produced more code without more finished software. Our move from an embedded BI tool to an in-product open-source stack went the other way: pages under a second, 3–4x the daily page views, and a lower cost to run. The difference was the work done before an agent wrote a line.
-
Why Enterprise AI Fails at Verification — and How I De-Risk It
The real work in LLM tagging isn't the model. It's building something to measure against, and putting the model's decision on rails so its judgment is both reliable and verifiable.
-
Proving It Holds Up: Turning Product Usage Into Statistically Defensible Impact Evidence
What it takes to move from dashboard correlations to evidence that survives third-party scrutiny — and why predictive analytics without measurement rigor won't win renewals.
-
Replacing Scattered Dashboards With One Analytics Cockpit
How consolidating fragmented BI into a self-hosted app on a refactored warehouse schema cut cost, built trust, and became the platform multiple business units actually adopted — a warehouse refactor nobody sees but everyone uses.
-
A Data Team Without a Roadmap Is a Ticket Queue
Data teams drift into reactive request-taking without one — and most data roadmaps are a stale slide nobody believes. A lightweight, horizon-based roadmap, kept honest by a weekly reconciliation between planned and actual work, is what turns scattered effort into a strategy leadership can see and trust.