Build and run, not just advise
Assessments are becoming a commodity. The hard part is what happens after the recommendation. We implement the systems we propose and then operate them, so the value shows up in your business rather than in a document.
Tokens are the unit of intelligence.
We do not just advise. We diagnose, build, and run.
An AI-native consultancy that finds where AI creates real leverage in your business, builds the agentic systems and tooling to capture it inside your existing workflows, and brings your people with us -- with your data kept sovereign.
Broadly applicable across industries and scale, run on our own agentic delivery capability. The proof: we run our own business this way.
What we are
Four things we do differently
Efficiency of intelligence
Tokens are the unit of intelligence, and we optimise for leverage per token. We measure what our own delivery costs to produce, which means our pricing is anchored to measured cost rather than billed hours.
Adoption is the real constraint
More AI programs fail on adoption than on technology. Change and education are a first-class part of the work: honest framing about what changes for people, and a program tuned to the issues your own teams raise.
Your data stays yours
Process maps, org charts and interview material are sensitive. Your particulars stay in storage you control. What we carry forward between engagements are patterns, never your particulars.
How it starts
From first contact to running systems
- 01
Tell us about your business
A short structured intake: your industry, your size, the processes that matter, where the pain is, and what you have already tried.
- 02
We research before we talk
Your industry, its competitive adoption picture and its regulatory horizon, built on research we have already accumulated rather than started from scratch.
- 03
You get a readiness assessment
A first read of your position and the opportunities visible from the outside, with every assertion sourced, reviewed by a person before it reaches you. Free, and meant to be useful even if you stop there.
- 04
If it is worth going further, we go deeper
Structured interviews across your stakeholders, refined opportunity and risk mapping, and a costed program of work. Then we build it and run it.
Working together
Five ways in, in order of depth
You can stop at any step. Most people start at the first one and decide from there.
- 01Free
AI readiness assessment
A structured read of where your organisation stands and where the near-term leverage is. Generated from your intake plus current industry research.
- 02Fixed fee
Diagnostic engagement
Deep research into your business and industry, a structured interview program across your stakeholders, and a prioritised opportunity and risk map with scoped projects.
- 03Program fee
Change and adoption enablement
Tuned education workshops built around the issues your own people raised, plus adoption playbooks and an honest account of what changes for whom.
- 04Per project
Build
We build the agentic systems, tooling and integrations identified in the diagnostic, embedded in the workflows your teams already use.
- 05Monthly retainer
Run
We operate, monitor and tune what we built: model upgrades, quarterly risk and industry refresh, and continuous improvement.
Being straight with you
What the free assessment actually is
It is the opening stage of the same work a paid engagement runs, stopped at the point where it would need to talk to your people. It is not a lesser, separate product built to generate leads. That also means it is honest about its limits: it can read your industry and reason about the processes you describe, but it cannot see what only your teams know. What you get is a real starting position, not a teaser.
- Built from your intake and public research, not from interviews with your staff.
- Opportunities are framed as hypotheses to test, not as commitments.
- No pricing or delivery estimates -- those need the deeper engagement.
- A person reviews it before it reaches you. We are early, and we would rather be slow than send you something automated and wrong.