Maturity
Where an organisation actually stands on the five levels — measured, not self-declared.
Approach
Transformation ambition stops at a wall.
The maturity model is the ruler. The benchmark is the measurement. And the measurement keeps returning the same finding: organisations stall at the boundary between Orchestrated and Exponential — because the first step across it is a question about trust.
The published evidence
Adoption is everywhere. Returns are not.
88%
Use AI
Of organisations use AI in at least one business function.
Source: McKinsey
~95%
Pilots stall
Of generative-AI pilots stall before production.
Source: MIT
94%
No value yet
Of organisations see no significant return on their enterprise AI investments.
Source: McKinsey
Where organisations stand
Five levels, and a wall between two of them
You can orchestrate without trusting: build the coordination layer, shift to goal stewardship, and still require a person on every consequential decision. Trust is what converts orchestration into autonomy — and autonomy into the exponential.
1 Additive AI replaces manual tasks. Growth stays tied to headcount.
2 Multiplicative Humans collaborate with generative AI. Productivity gains are real, but legacy data creates drag.
3 Orchestrated The shift to goal stewardship. Central layers coordinate specialised agent workflows.
4 Exponential Trustworthy agent ecosystems operate with minimal intervention. Return is measured as value created, not hours saved.
5 OGI AI is the integrated brain. Humans and agents function as one synchronous organism.
What the benchmark keeps finding: organisations cluster at Multiplicative, and Orchestrated is thinly occupied. “Nothing in the middle” is not a mood — it has a location, and a cause.
How we work
From where you stand to a function that runs itself
Five layers, walked top to bottom — one vertically integrated route from discovery to operation, with the capabilities, the products and the partners each doing their part. One firm across the whole of it, not an adviser who leaves before the build.
03
Model your processes
The work as it actually flows, made visible and redesignable.
What the research measures
Three theses, one instrument
The benchmark is a research instrument, not a scoring quiz. Every completion adds evidence on why transformation stalls — and gives the participant a diagnosis in return.
The trust gap
Quality, auditability, security: what an audit function needs before it trusts an agent the way it trusts a person — and which of the three is the blocker.
The cost to beat
The economic threshold a process must cross: what the work costs run traditionally, against what it costs run agentically.
Proof
Running today, not piloted
Risk simulation runs live today with a global consumer goods company. Discretion is part of the offer: we let the work stand without naming the client.
Transformational impact comes from combining organisational design, process re-engineering and agentic technology — never from the technology alone.
Find out where you stand.
The benchmark is live, conversational, and tells you something before we ever speak.
Take the benchmark