Start Here
Where AI actually improves decision quality — not just output quality.
I help operators and founders judge where AI actually improves decision quality in messy, high-stakes environments. The frameworks below and the 1-page Scorecard are how I make that lens portable.
Free, ungated
The AI Decision Effectiveness Scorecard
A 1-page rubric that maps the IRDA framework against Decision Maturity. Use it to grade where AI sits in any internal project. No email required.
Download the Scorecard (PDF)If it's useful, the weekly newsletter is below — completely optional.
The three ideas to remember
-
1. AI Decision Effectiveness — not just Output Quality
LLMs are currently evaluated on text generation. Operators need them evaluated on decision leverage. The right question is rarely "did the model produce something readable?" — it's "did the right decision actually get made downstream?"
-
2. The IRDA Framework
Information → Recommendation → Decision → Action. The four stages where AI can sit in an enterprise workflow. Most "AI in production" stories live at I or R. The interesting (and unsolved) work is at D and A. Read the full framework →
-
3. The Messy-Data Advantage
Data friction (healthcare pricing, MRFs, internal taxonomies) is a moat. Polished demos fail in production because production is messy. ClearPriceHealth is the domain proof for this idea — healthcare price transparency forces every IRDA stage into the open.
Weekly notes from the lens
One issue per week. The canon applied to one news event, one field note, and three things worth bookmarking. Reply to any issue — replies are read.
Next stops
- The frameworks — IRDA, Decision Maturity, how they compose.
- The journal — longer essays organized by tier (Individual → Ecosystem) and topic.
- Field notes — shorter, messier reports from real production work.
- In Practice — what I’m building and learning, with the framework in production.