Assess: AI inventory and maturity, straight from the code.
Most organisations cannot say how many AI assets they run, let alone how mature each one is. Assess reads the repositories, discovers every model, integration, agent and call site, scores the project on six maturity lenses, and ranks the concrete steps that would raise the score most. It is the instrument behind the Assess phase of AIM.
The Assess phase, measured from the code rather than the interviews.
An assessment built on workshops tells you what people believe is running. Assess reads what is actually running: which models are called from where, which agents exist, which configurations are live, and which of it is used at all. The maturity score is computed from evidence in the repository, and every recommended step points to the places in the code it applies.
Where a project stands, and what would move it.
Each lens is scored from evidence. Pick one and Assess shows what is going well, the next steps, and the lift each step is expected to add.
Observability
Are calls traced, are failures classified, is cost attributed?
Governance
Are intended uses documented, are keys separated, is there an approval path?
Security
Are tools scoped narrowly, are secrets managed, are inputs guarded?
Engineering
Is provider and model selection centralised, is there a test suite for behaviour?
Data
Is retrieval governed, is sensitive data classified, is lineage recorded?
Lifecycle
Are models versioned, are evaluations repeatable, is there a path to retire?
Five screens from a live assessment.
From the portfolio sorted worst first to the ranked list of what to fix, this is what an Assess engagement produces in its first days.
Live captures from October 2026. Project names shown are from a pilot environment.
The first three weeks of AIM.
Assess runs in the Assess phase and feeds the ranked backlog that Implement starts from. The score gives the sponsor a baseline; the ranked steps give the pod its first fortnight of work.
Tell us the number you need to move.
A 45-minute working session with an operator who has run the kind of work you are describing. You will get an honest read on where your programme stands and what it would take to move it.