Senior enough to know your business. Hands-on enough to know what AI really does.
How we work, step by step, and why the controls are part of the work rather than a review at the end.
Senior and hands-on
Consulting sells pyramids: a partner to sign, a manager to coordinate, and the work done by people learning on your time. We do not. The person who scopes the work is the person who does it.
And we build. In AI there are no experts, only people who experiment more. We run agents on our own business every day, which is why we can tell you what a demonstration hides.
The five steps
- 01
Find
What we do: with your business heads, we go through the work your US entity actually does and find where AI pays off. We size each candidate in money and time.
What you get: a shortlist with a business case per use case, and the ones to drop.
Compliance designed in: each candidate carries its risk class and the controls it will need, before anyone builds.
- 02
Design
What we do: we redesign the work itself: what the AI does, what people keep, who owns each decision, and what always goes to a person.
What you get: a target process, decision rights, and the measure it will be judged on.
Compliance designed in: reserved decisions and escalation paths are written into the design, in plain language the process owner signs.
- 03
Select
What we do: we select the best builder for each use case: your group IT, a vendor, a specialist firm, or us where a pattern we already run fits. We take no referral fees.
What you get: a comparison on evidence, and a recommendation you can defend.
Compliance designed in: every candidate is asked for the evidence its system can produce, not for promises.
- 04
Deliver
What we do: we lead delivery with the builder you chose, against the agreed metric, in your environment.
What you get: a system in production, and the dashboard that shows how it behaves.
Compliance designed in: controls are built and tested layer by layer as the system is, not retrofitted.
- 05
Adopt
What we do: we change roles, routines and training so the gains last beyond the project.
What you get: the work running the new way, with the people who do it trained for it.
Compliance designed in: autonomy is granted step by step, on evidence from the monitoring, never at go-live.
Proven in our lab first
Our lab, plzfix.ai, runs agents on our own consulting work: market watch, prospecting, proposal drafting, structured interviews, document production, recruitment.
When it fits, we can bring a pattern we already run, with its prompts, its configuration and its controls. Otherwise we select the builder that fits best. The point is not to sell you ours; it is that we have learned what breaks before you pay to find out.
Further reading
- Compliant by Design — Our paper on building generative AI systems the business can control.
- From prompts to org chart — How companies become AI-native, published by our lab.