Method
How we work
Every engagement is fixed in scope, duration, and fee, produces named artefacts, and ends at a decision gate. You should always know what you are buying and what evidence will be on the table when it ends.
- Commercial shape
- Fixed fee, fixed duration, named deliverables
- Team
- The people on the call are the people on the work
Principles
Six commitments we can be held to
01
A baseline before a target
We write down what the process costs and how long it takes today, before anyone proposes a percentage improvement. A target without a baseline is a wish with a number attached.
02
Decision rights before deployment
Approve, override, and escalate are assigned to named people on the specific workflow, not to a committee, and not to a policy document nobody has read.
03
Small, consequential, measurable
The first case must matter enough for a leader to defend it and be small enough to finish. Cases that fail either test go later in the sequence.
04
The failure path is designed first
What happens when the model is wrong, the vendor changes the model, or the volume triples. If the exception path is undefined, the workflow is not ready.
05
We hand it over
Every engagement ends with a runbook, an evaluation harness your team can run, and a session where your people operate the thing without us in the room.
06
We say no in writing
When a request is not ready, or is better solved without AI, we write that down with the reason. It has cost us work. It has also kept clients longer.
The ladder
Proof, Product, Platform
The commercial ladder runs beside the NATIVE loop. Each rung has its own exit criteria, and none of them assumes the next.
Proof
2 to 3 weeks
A working prototype that makes a decision possible
Gate: A value map exists: metric, named owner, expected benefit, validation method. If it does not, the engagement stops here and that is a legitimate outcome.
We refuse: A prototype with no named decision attached to it.
Product
6 to 12 weeks
One real application carrying real work in production
Gate: The pilot charter's pass line was met on a fully loaded cost basis, in the real workflow, on the hardest cases and not only the clean ones. Adoption, quality, and risk all cleared.
We refuse: Production deployment without an escalation path and a kill switch.
Platform
2 to 4 quarters, in sequenced waves
A governed capability that lets many teams build safely
Gate: The first wave is running in routine operations with monitoring, an owner, and a documented crossover point.
We refuse: A rollout with no named owner per department.
Commercial policy
The rules we hold ourselves to
- Fixed fee, fixed duration, named deliverables.
- No hourly billing. No time-and-materials creep. No change order for a question.
- You own everything from day one: code, prompts, evaluation sets, calibration data, and documentation, in exportable form.
- A stage that ends in a decision not to proceed is a completed engagement, not a failure.
- We do not sell seats and we do not compete on hourly rate.
FAQ
Questions about working with us
- Do you take engagements without a diagnostic first?
- Yes, when a client already has a written baseline, a named owner, and a metric they will be held to. When those are missing we say so rather than billing for a build that cannot be judged.
- Who actually does the work?
- The people on the first call are on the engagement. We do not run a pyramid where senior people sell and juniors deliver, and we will name the team before you sign.
- What happens if the evidence says stop?
- The engagement is complete. A stage that ends in a decision not to proceed has produced the thing you paid for: a defensible decision, made earlier and more cheaply than the alternative.
- Do we own what you build?
- From day one. Code, prompts, evaluation sets, calibration data, and documentation, in exportable form, with a handover session and a runbook.
Next step
Start with a baseline, not a proposal
Take the NATIVE Audit in fifteen minutes, or book a scoping call and bring the pilots you already have running.