A methodical read of where the organisation stands
Seven dimensions, forty-nine areas, scored identically in every engagement.
Five levels of AI-nativeness
Not a rating but a description of five companies — the same business at five degrees of AI-nativeness — so the next version of yours is something specific. Most sit at 2 or 3, and the return is released between 3 and 4.
-
Work is done by hand.
-
Drafts and summaries. A few individuals, no consistent practice.
-
Prepares material. A person finishes and decides. Whole company, individual work.
-
Takes a task end to end. A person approves. Selected processes, wired into systems.
-
Runs processes and coordinates between them. Across departments, end to end.
Below the threshold
The model operates without the company's context. It cannot determine which price list applies, which of two procedures is followed in practice, or what an exception is. Every output is co-authored by a person, and hours remain broadly unchanged, because the review is the work.
Above the threshold
The model operates against documented rules, so it finishes the task and a person approves it. Capacity stops being a function of headcount, and the organisation starts to look like the one on the right of the scale rather than a faster version of the one on the left.
Seven dimensions, forty-nine areas
Strategy
- Vision & ambition
- Portfolio & prioritization
- Investment & funding model
- AI in the offering
- Commercial model
- Market sensing
- Strategic agility
Operations
- Process redesign
- Automation coverage
- Human–AI collaboration
- Decision architecture
- Customer experience
- Operational instrumentation
- Process ownership
Technology
- Tooling stack
- Build & integration
- Architecture & scalability
- Deployment velocity
- Reliability & operations
- Evaluation & QA
- AI security
Data
- Data quality & trust
- Accessibility & integration
- Knowledge management
- Information architecture
- Data governance
- Context readiness
- Proprietary data
People
- Leadership capability
- Workforce AI literacy
- Adoption depth
- Talent & retention
- Role & work design
- Culture & safety
- Learning velocity
Governance
- Policy & acceptable use
- Risk management
- Regulatory compliance
- Responsible AI
- Accountability & oversight
- Vendor & model governance
- Incident response
Value
- Measurement discipline
- Cost efficiency
- Productivity leverage
- Revenue & growth
- Unit economics
- Customer value
- Differentiation
Facing what is there
Scoring forces the organisation to say out loud how it prices, what an exception is and who decides — and the gaps that surfaces are usually older than any AI project. That is uncomfortable and it is the point: this is the first time there has been a concrete reason to finish it.
A sequence, not a score
For each area, the description one level above the current score defines the next step. The lowest-scoring dimensions determine which processes are implemented first — and the levels themselves describe what the company looks like once it has moved.
How a level is climbed: five steps, then again
The same five steps, whether one process or fifty.
-
1
Selection
Which process moves first, and why that one. We open with the index or a short discovery, then record the baseline before anything is touched, so there is something to measure against at the end.
-
2
Codification
The rules the process stands on: pricing logic, exceptions, decision authority and hard limits, documented with the people who hold them. This part stays with you and carries into every process after it.
-
3
Implementation
Built inside the tools your team already runs, with a person approving anything that leaves the company. In production within thirty days, not as a pilot environment but as the way the work is now done.
-
4
Adaptation
Roles redesigned around the changed process, and the team trained on the live system using their own cases rather than a demo. A one-page standard states what the system does, what the person does, and when to escalate.
-
5
Development
Measured against the baseline recorded at the start, corrected, and kept current as prices, rules and people change. Tools are swapped without a rebuild, because the architecture was never tied to one of them.
One process, selected
We open with the index or a short discovery, agree which single process moves first and why that one, and record the baseline before anything is touched.
Its rules, written down
Pricing logic, exceptions, decision authority and hard limits, documented with the people who hold them. This part stays with you and carries into every process after it.
Live, and measured
Built into the systems you already use, the team trained on it, and the effect compared against the baseline recorded at the start.
Let's write down how your company actually works.
- 45 minutes, and no preparation needed
- An honest read on where you sit today
- The first process worth automating — and why that one
