In previous articles, I’ve argued that transformation should be defined by quality (i.e. meeting customer’s needs) rather than good (a subjective projection of quality), that cost reduction and efficiency are not the same, and that learning occurs when systems develop the capacity to sense and adapt.
This article asks the next question: once a system can learn, how do we measure whether it is learning well? Is it adapting towards its purpose or is it drifting from the patterns it is meant to reinforce?
The Limits of Maturity
For decades, maturity models have been the default instrument of transformation. They assign scores, define levels, and imply that organisations progress linearly from “ad hoc” to “optimised.”
Useful, up to a point. But in practice, maturity models reward consistency, not adaptability. They tell you how aligned your current practices are to a static ideal - not how well the system adjusts when the environment changes. In Ashby’s book, “A Design for a Brain” (1952), he notes that in complex systems, stability is not achieved by resisting change but by absorbing it.
What leaders need now is not a ladder of compliance, but a compass of coherence.
Coherence as a Measure
Coherence is the degree to which an organisation’s purpose, structure, process, and behaviour reinforce each other under changing conditions. Clients have often said to me that transformation creates noise. Coherence is, both figuratively and mathematically, the opposite.
Unlike maturity, coherence can rise or fall daily. It is not a level achieved but a state maintained.
In WorkLattice AI, coherence is represented as a dynamic graph: each node a role, system, process or policy, each edge a dependency. As changes propagate through the lattice, the platform measures how much variance emerges between declared intent and actual behaviour.
Measuring the Invisible
Traditional metrics capture outputs (speed, cost, compliance). Coherence metrics capture relationships:
Alignment — how purpose and performance connect.
Resonance — how well information and intent flow through the system.
Resilience — how changes in one part of the system affect others.
Drift — how quickly a process diverges from its intended design.
Redundancy — whether duplicate requirements emerge under pressure.
WorkLattice’s model tests these continuously.
When a policy change increases alignment but reduces resilience, the system flags a trade-off and simulates alternatives.
When coherence is high, investments in technology, process, or people multiply rather than fragment.
When coherence is low, the same investment diffuses through friction, redundancy, and misaligned incentives.
A coherent organisation can flex without fragmenting. Its decision rules, incentives, and capabilities are internally consistent and externally responsive. A coherent organisation can absorb disturbance because its interconnections are visible and well designed.
Understanding the system is what turns spending into strength.
From Scorecards to Signatures
Every organisation develops a distinctive coherence signature, a pattern of how it holds together under strain. Rather than a static benchmark, coherence signatures are used for comparison across time and context:
The same enterprise under regulatory pressure vs. under growth pressure.
The same function before and after automation.
The aim is not to rate the organisation but to understand its structural rhythm, how quickly and cleanly it realigns itself when perturbed and how it compares to other organisations that have similar mission or purpose.
The Role of AI
This kind of continuous, system-level measurement would be impossible manually. AI makes it feasible by acting as the instrument:
Continuously mapping flows of work, data, and decision rights.
Detecting when patterns diverge from intended design.
Re-running small-scale simulations to estimate impact before leaders feel it.
But AI doesn’t sit outside the system measuring it. It sits inside the system participating in it.
If the system is coherent: AI amplifies alignment, clarity, and learning.
If the system is incoherent: AI amplifies drift, contradiction, and unintended consequences.
AI is only as intelligent as the system it is interpreting.
Without coherence, even the most advanced model will misread signal as noise and noise as signal.
This is why tools such as WorkLattice focus first on mapping and modelling relationships: flows of work, policy dependencies, decision rights. Because the AI can only learn what the organisation already knows about itself.
Seen this way, AI doesn’t replace organisational intelligence, it gives leaders a nervous system capable of perceiving the system they are responsible for.
Otherwise: AI becomes a way to get lost very fast.
Looking Ahead
In Article 1, we reframed quality.
In Article 2, we examined the genesis of work.
In Article 3, we explored system learning.
This article introduced coherence as a measure of that learning in motion.
Next (Article 5), I’ll explore how to design for coherence: the architectural choices that allow systems to stay flexible without losing form.
Reflection Question
If AI will amplify whatever system it enters, are you confident your organisation is coherent enough for that to be an advantage rather than a liability?