What we say before signing, not after.
How engagements work, what is guaranteed, what we do not do and who we are not right for. All of this is said here so that it is not a surprise in the proposal.
The three services
The same client goes through all three.
Vanguard, Forge and Taskforce are not three kinds of client: they are the cycle the same client goes through. They are contracted separately and quoted separately. A verdict of ours never arrives with a quote to execute it.
Vanguard
A shift is reaching you and you have to decide what to do about it before committing budget. What comes out is a criterion you can defend the decision with, not a presentation.
Build · “I want to implement”Forge
What you decided has to be made real. It gets built and handed over working, with no future dependency on us, fully owned and controlled by you. In production, not in a sandbox.
Sustain · “I want to stabilize”Taskforce
Something critical is at risk or has just failed. It gets diagnosed, the operation gets recovered and your team is left ready for the next time, with someone who has already run those recoveries.
The guarantee
You do not pay by the hour. You pay for the result.
Fixed price
Never by the hour. The fee is anchored against the consequence of the decision, not against how long it takes to resolve it. Price and timeline are set after the 25-minute call, in writing, before starting. They do not change during the engagement.
Acceptance criteria
Written before starting. You know from the outset what has to be on the table to call the work good.
Correction at no cost
If the deliverable does not meet that criterion, it gets corrected. If after the correction it still does not meet it, you do not pay the balance of the engagement.
Phases with a cut
Each phase is accepted before the next one starts and at each cut you decide whether to continue.
What is guaranteed and what is not. The result of the contracted work is guaranteed. The commercial result of your company is not, since it depends on decisions we do not make.
The scope. How far the guarantee reaches is set out in writing in the proposal, before starting.
Who delivers
Whoever sells is whoever delivers.
Senior is not sold to deliver junior. The founder leads and delivers every engagement, with backup from senior colleagues when the specialty demands it.
The person who hears the problem is the same one who signs the verdict and the one who is there on cutover day. Whoever backs them up is named in the proposal. Everything produced is yours from day one, in your own repository. Who delivers.
Independence
Whoever builds does not validate.
The firm that implemented it
It bills the build. It takes a commission from the vendor it chose. And it signs the report that says it works.
the same one
The firm that validates it
It does not resell technology. It takes no commission from anyone. It is not going to build what it has just judged.
When the firm builds something, the independent review is contracted and paid by you, with whoever you choose. We give access to everything and we do not review ourselves. The rule works both ways: that is why a verdict of ours never arrives with a quote to execute it.
And a diagnosis can conclude that there is no case. When that happens, the diagnosis is charged, the reason is explained in writing and the project is not done.
The rule is one of sequence, not exclusion: the firm can guide, build, or validate what someone else built, but never build and validate the same system. It is the same logic as the Chicago Principles from AIQA (May 2026) and the notified body model in the EU AI Act: self-attestation is not governance.
The frontier
At the edge there are no best practices.
A best practice is a pattern that worked enough times for somebody to write it down. That makes it an artifact that arrives late.
Where the movement has just started nobody has done it enough times yet, so there is no pattern to copy. What is proven is something else: having reached the right decision the first time, with no safety net, with money and continuity on the line. That is not learned by reading what worked for somebody else.
The limits
What this firm does not do.
It does not place staff or sell hours.
No dedicated teams. That is capacity and what is sold here is judgment.
It does not resell technology or take commissions.
No license or partnership should be able to bias a piece of advice. It is what makes the word independent verifiable.
It does not work as a subcontractor for integrators.
You cannot audit on Tuesday the work of whoever bills you on Thursday.
It does not give generic technology training.
The education that is done is about how to decide, not about how to use a tool.
It does not stay operating your infrastructure.
It gets stabilized, documented and handed back. A permanent dependency contradicts the promise.
It does not deliver with juniors.
If the work cannot be done by someone senior, it is not work for this firm.
Who it is for
A moment defines it, not an industry.
Three conditions are checked before accepting an engagement. They qualify, they do not segment: a small company can meet them and a corporation can fail to meet them.
- There is a frontier move already touching the company, not a curiosity or a topic of conversation.
- There is someone with authority and budget who has to decide, not an interested party without a mandate.
- The stakes are real: there is something at stake, whether money, continuity, regulation, reputation or opportunity.
Every client receives the same degree of seriousness, rigor and confidentiality, whatever their size, country of origin or standing.
Who it is not for
Whoever needs to add people, not to solve a problem. If the team already knows what to do and what is missing is capacity, that gets bought by the hour. Here the engagement is always a deliverable that ends, even when it has to be built or sustained.
Whoever is looking for the industry best practice. If yours already has a proven answer, copy it: this firm is not the one needed.
What is the decision you have in front of you this week?