Design partners
Deploy one governed workflow. Shape the substrate.
A design partnership is how the substrate enters production before it is a list-price product. It is paid, bounded, and evidence-producing: we deploy one high-value governed workflow inside your perimeter, with a dedicated engineer, and you shape the platform as it productizes. Not a pilot that goes nowhere, and not an open-ended consulting engagement.
§ 01 What you get
A workflow in production, and the evidence to defend it.
The point of a design partnership is a real outcome in your environment, not a trial license. You leave with a governed workflow running and an audit record an examiner can interrogate.
A deployed governed workflow
We stand the substrate up inside your perimeter and put one high-value workflow into production under governance, with a dedicated engineer who onboards your operators and stays through enablement.
The evidence your examiners need
Every agent action attributed to a specific agent under a specific policy, hash-chained and verifiable from any point. When your auditor or regulator asks what the agents did, you have the record, not a reconstruction.
Influence on the platform
Your real requirements shape the roadmap. Capabilities you need become product capabilities, not one-off custom code. You are early enough that the platform bends to your environment.
Design-partner terms
Bounded scope, a defined end state, and pricing structured for an early partner rather than a list-price customer. The relationship is built to convert into a platform license when the evidence supports it.
§ 02 What we get
An honest exchange, stated up front.
A design partnership is a trade, and we would rather name our side of it than pretend it is charity. Your confidential data and workflows stay yours throughout.
A real production validation
One genuine deployment in a real environment is worth more than a hundred demos. It proves the substrate where it matters and tells us what enterprise production actually requires.
The right to generalize
We retain the right to turn non-confidential learning from the engagement into reusable platform capability. Your confidential data and workflows stay yours; the general patterns make the product better for everyone.
A reference, where permitted
When you are willing and able, a reference or case study. Never required, never assumed, and only what you approve.
§ 03 The motion
Evaluation, to bounded pilot, to design partner.
Three stages, sized to your commitment. Each has a defined success criterion and a clear exit, so you are never locked into the next stage before the last one earns it. The pilot gate is a disproof attempt, not a success demo.
Evaluation
About 4 to 6 weeks
We assess where governed AI agents fit your operations and map your governance posture against agentic-AI requirements. You get an independent read. No platform commitment.
Pilot
About 12 to 16 weeks
We stand the substrate up against one bounded workflow in your environment, with success criteria agreed up front. Both sides have a clear exit.
Design partner
18 to 36 months
The substrate is in production on your real workflows, with a dedicated engineer through enablement. You are the operating reference.
§ 04 Who fits
We onboard a few each quarter. Here is the fit.
Because every partnership gets a dedicated engineer, we take a limited number at a time. A strong fit looks like this:
- 01A material governed-AI problem that is already costly or blocking a deployment.
- 02An executive sponsor who can authorize a pilot, and a technical owner who can deploy.
- 03A risk or control owner who needs attribution, reconstruction, approval, or recovery.
- 04Access to a real workflow and the evidence around it.
- 05An environment the substrate can run in, including in-perimeter or air-gap-capable deployments.
- 06Willingness to give structured feedback and a defined budget.