Shipping Better MLPs Smarter
A governed architecture for turning open-ended AI planning into bounded, reviewable product work.
Read article →NBM4 / THE RESEARCH ROOM
Turn product uncertainty into a system that learns.
A practical operating framework for discovering, validating, building, and growing products—without pretending outcomes are deterministic.
Explore the framework ↗THE OPERATING FRAMEWORK
Two phases. Four stages. One evidence loop.
NBM4 connects early opportunity discovery with systematic engineering delivery. Each stage produces evidence for the next—and provides explicit falsification signals to pivot or iterate before committing engineering capacity.
THE PREMISE
Products are hypotheses about future behavior.
Code can execute deterministically. Product outcomes never did.
NBM4 gives teams a common structure for managing that uncertainty: explicit assumptions, typed product artefacts, evidence gates, and learning loops that keep strategy connected to delivery.
Read the NBM4 introduction →LATEST THINKING
Research without ceremonial certainty.
The Sprint Refinement Session Nobody Wins
Why market responsiveness and engineering capacity collide—and why better facilitation cannot fix the system.
Read article → 03 Strategy
Product Strategy That Learns in HADI Cycles
Wiring strategy, evidence, and NBM4 stages into an operating model that can actually change its mind.
Read article →
25+ YEARS IN THE FIELD ABOUT THE AUTHOR
Built in the field, not assembled from framework-shaped quotes.
Michael Kozloff has led product management, business development, and transformation work across Microsoft, VMware, IBM, startups, and more than twenty products. NBM4 is the operating framework extracted from that work—and continuously revised by new evidence.
Meet Michael →START WITH THE SYSTEM