Start with an hour. Scale only if the math holds.
The practice prices where agents run, not whether you buy seats or devices. Both of those are table stakes; we never argue against either. Every engagement below runs on the same six-evaluator framework and the same engine that powers the public model.
The executive briefing
One hour, no cost, no deck theater. We walk your leadership through the category, the placement question, and the three-baseline math, against your own scenario if you bring the numbers. You leave knowing whether this is worth pursuing, including when the honest answer is that it is not.
The Envisioning Workshop
A working session with your architects and finance owners. We score your actual workloads against all six evaluators, run the placement model live on your fleet numbers, and leave you with a written placement map: what runs locally, what stays in the cloud, and why. The math on the public model page is the part of this workshop you can run without us.
The Build engagement
When the workshop says a workload belongs on your hardware, we build the proof: a working local-AI implementation on your silicon, wrapped in the governance your IT organization already operates. Reference blueprints from the public library are the starting point, adapted to your environment and your control framework.