One operating model for product, IT, data and AI.
McKinsey, BCG and Gartner now describe the same target architecture from three different angles. We've turned it into a working operating system. Four layers, three containers, three lenses, one source of truth.
Three vantage points, the same operating model
Independent of each other, published months apart, in 2026.
Five pillars, one integrated system.
Business model, operating model, governance, talent and culture, technology and data. Read as one system, not five compartments.
10-20-70.
10% on the algorithm, 20% on data and tech, 70% on people and processes. The work sits outside the model.
PPM is now SPM.
Project & Portfolio Management was retired as a category. Strategic Portfolio Management replaced it. Different first question.
From strategic claim to running experiment
The architecture reads top-down. One bet hosts opportunities. Each opportunity drives initiatives. Each initiative resolves through experiments.
Bet
A falsifiable strategic claim, dated. What we believe strongly enough to commit capacity to, and accept seeing invalidated.
Opportunity
A customer need, problem or business lever that contributes to the bet. Drawn from Teresa Torres' Opportunity Solution Tree.
Initiative
The solution explored to address one or more opportunities. What product and engineering teams carry day-to-day.
Experiment
A concrete test of one hypothesis. Can succeed, fail, or stay inconclusive. The one level where invalidation is explicitly accepted.
Where every bet finds its level
Bets live at the product level. Not at the portfolio level (too broad). Not at the initiative level (too narrow). One rule, a lot less ambiguity.
The whole organisation.
Coherent grouping by BU, market or mission.
A team, a scope, a bet.
One source of truth. Three readings.
Same data model. No re-entry. No reconciliation. Each role asks a different question and the system answers it directly.
“What are we learning this week?”
- The Opportunity Solution Tree, live
- Experiments by hypothesis and outcome
- Weekly invalidation, not weekly status
Click a lens to switch view. The underlying data does not move.
Observability that the comex can actually read
Four layers stacked. The bridge between MLOps and strategy: each technical signal climbs back up to a bet.
Lifecycle
Where every use case sits across the six phases. What's blocked, for how long.
Compliance
Which use cases qualify as high-risk under the EU AI Act. Conformity assessment, human oversight, documented.
Technical quality
Input data quality, model drift, performance decay, operational alerts.
Value
What business value each use case actually captures, tied to the bet it serves.
Portfolio operating system.
Not another PPM, not an AI governance layer bolted onto product. A single source of truth on bets, opportunities, initiatives and experiments, read through the PM, PMO and CDAIO lenses, anchored to a watchtower that ties value back to bets.