How we work
Train. Build. Ship. One arc, ninety days.
We are practitioners. We do three things in this order, with your people, on your processes, against your numbers.
- 01~ 1.5 days
Train.
Your leaders see what is possible.
- 02~ 2 weeks
Build.
Your team builds real agents on real workflows.
- 03~ 90 days
Ship.
The first agents go into production.
- 04forever
Stays.
The capability stays with your people.
Why this order
Leadership first. Always.
Every serious AI rollout we have been part of started with a room that had the CEO in it. Every stalled one had a leadership team that had not touched the tools.
Budget, permission, adoption, focus: all of it flows from the top. If your CEO sends the CTO to “check this out,” we tell them to pick a different vendor. We are not interested in engagements that bypass the room where the decisions get made.
Once the leadership team has seen what is possible and named the first action, the rest of the arc is a build problem, not a buy-in problem. That is the order. It is not the order most consultancies follow. It is the only order that ships.
Train
The Leadership Accelerator.
A 1.5-day working session for your CEO and direct reports. They walk out with a named first action, in their function, on their actual processes.
The room is the CEO and the eight to twelve people who actually run the company. No proxies. No skip-levels. Outside of the CEO, every leader brings one process they own that is currently broken or slow.
Day one is what AI can actually do today, demonstrated on those processes by a practitioner from one of the labs building AI. Not a deck. A live build. By lunch on day two, every leader has a 30-day commitment in their own words. By the end of day two, the CEO has picked the function we go deep on next.
At Constance, this room produced eleven commitments. Eight shipped within sixty days. At a global telecom, the same format took the leadership conversation about AI from quarterly slides to weekly working sessions. The arc starts here.
Build
The Functional Accelerator and Agentic Fest.
A two-week sprint. Your best people, our practitioners, your sandbox. Demo day at the end.
We pick one function from the leadership session and put eight to twelve of your people in a sandboxed VM with practitioners who have shipped AI in production. They build the agents the leaders said they wanted. Real workflows. Real data, scrubbed. Real demos.
We do not pretend this is training. It is build work, and the people in the room ship. Every team commits to one agent on day one and demos it on day ten. The Agentic Fest at the end is a demo day for the rest of the leadership team. The agents that survive demo day go into the next stage.
“The fest format produces real working software in two weeks. Most consulting engagements produce a roadmap.” Deloitte partner, after observing the e&PPF Prague fest
Ship
The Pod.
A 90-day pod: your best people, our practitioners, one shared backlog. We leave when your team can run it without us.
The pod takes the agents that survived demo day and puts them into production against real usage. It is the first time most companies see one of their AI experiments actually serve business traffic.
We staff the pod with two of your best people and one of our practitioners. Ratios shift toward your people week by week. By the end of 90 days, the pod is yours. The practitioner rotates out. The agents keep running. The capability stays. That is the only outcome we accept.
What we do not do
A short, honest list.
- Six-month strategy decks
- Transformation-as-a-service contracts
- 200-slide diagnostic readouts
- Replacing your people
- Engagements without a CEO in the room
- Outcomes priced by hour
We are not a strategy firm. We are not a coding bootcamp. We are not a dev shop. We are the team that gets your people to ship the first agents and then leaves.