Quarterly
Spring 2026
A threatcasting read on the agentic year ahead.
Four domains, three anomalies, four threat models, six moves, a twenty-four signal registry.
Everyone is buying AI capability faster than they can turn it into return.The winners will be the ones who build the operating model that converts capability into alpha, and most never will.
40%+
of agentic AI projects are forecast to be canceled by the end of 2027.
Gartner · 2025
70%
of transformations fail to meet their objectives.
Kotter · BCG · McKinsey
17%
of large technology programs run so far over that they threaten the company's existence.
McKinsey & Oxford
Scroll
A turning point, in the Perez sense
The capability curve went vertical. The operating model did not.
Every technological revolution runs the same arc. A new capability arrives, capital floods toward it faster than the institutions that would govern it can adapt, and a frenzy inflates. We are inside one now.
In the stretch between installation and deployment, the capability curve goes vertical while the operating model, the governance, and the organizational form move at institutional speed. This is the pattern Carlota Perez traced across two centuries of surges. Inside a single company it shows up as a widening gap. The capability can already do the work, and the P&L stays flat, because the system around it takes far longer to change. A capability is a purchase. An operating model is structural work. The return sits in that gap until the structure catches up, and the organizations that pull ahead are the ones that close it first.
Installation and deployment across surges
A declining surge is displaced by a new one, two sigmoids meeting at a turning point. The new surge moves through disruption and frenzy in installation, then synergy and maturity in deployment.
Insight · enterprise risk
Orienteering the Spectrum
Reading enterprise risk in the frenzy, when the capability curve outruns the institutions built to govern it. Introduces PRISM.
Read the piece →
Insight · constraints
The Fitness Trap
The organizations that optimized hardest for the last cycle are the most brittle in the next. When optimization becomes the constraint.
Read the piece →
Re-deriving the operating model · Outcome-Owned Architecture
OOA re-derives the operating model from the decision up.
It starts from a plain question the old model never asked: which decisions should software own, which stay with a named person, and where do the two meet. That line is the operating model. Draw it right and the capability has a shape that lets it pay.
9.2×Portfolio benefit to cost · ρ = V / (C + R)
Descend into the practice →The decision loop, partitioned by the judgment boundary
A cybernetic control system. The machine owns the loop inside declared policy, humans own the three loops that govern and interpret it, and ownership transfers only at the seam. Expand for the full figure.
Sensing · strategy · operations
Advantage in this market comes from agility, and agility is built across sensing, strategy, and operations.
In the frenzy, capital rushes to acquire the capability. The harder question, how it operationalizes into P&L expansion, gets deferred to later, and later never comes. Advantage goes to the organizations that answer it. We work in all three practices that answer it. OOA is the one we are publishing in full, and the depth behind the other two is the same.
01
Strategic Foresight
See what is coming, and discern what matters, before the signals force a decision you did not prepare for.
Continuous behavioral sensing wired into the decisions that matter, grounded in the intelligence cycle and forty years of scholarship.
Enter the practice →
02
Adaptive Strategy
Decide what to do under real uncertainty, when the plan cannot see the whole board.
Strategy as choice and emergence held together, tested against evidence rather than appetite.
Enter the practice →
03
Outcome-Owned Architecture
Build the operating model that turns the decision into return, in the agentic era.
Re-derive the operating model from the decision up, so capability has a shape that lets it pay.
Published in full
Enter the practice →
The shared lineageAll three are derived from the same foundational theory: Theory of Constraints, Lean, and Socio-Technical Systems. The full architecture, theory to practice to instruments, is on Methods.
Current work
What we are working on now.
Featured essay
Assessing the Unbearable Lightness of the Symbiotic Enterprise
Eight assumptions in the new vision of the AI-reinvented enterprise, read against the operations science of execution.
Read the essay →
Reflection
Monthly
Where the pilot count stops being evidence.
Working paper
Monthly
One named owner, and what the loop costs without one.
We will tell you where it does not pay, and we will tell you to walk away.
Request a briefingOr read the thinking behind the method, and the record behind it, on About.