Aller au contenu
The Frontière method

What we sell, in one equation.

Three terms, none of them optional. Remove the engine and what remains is a classic consultancy — limited by hours. Remove the matrices and what remains is a flood of analysis — with no reading. Remove the partners and what remains is software — no one to answer for the conclusion. The whole equation is what you buy.

Tesuji

The engine.

Frontier agentic systems, designed engagement by engagement — AI design is our expertise.

+
The matrices

The precision.

Goban, our repertoire of proprietary conceptual matrices — one per family of decision.

+
The senior partners

The human touch.

The framing, the interviews, the contradiction, the conclusion — committed and signed.

=

A decision investigated, contradicted, signed —

delivered within the time of the decision

01 / The engine — Tesuji

Agentic systems, designed for a single question.

Tesuji — in Go, the skilful move. Our engine is not a generic tool you query: it is a chain of agents assembled for your decision. Designing it is a craft — AI design: which sources, in what reading order, under which grids, with which controls.

Designed per engagement

The architecture fits the question. A team assessment, a sale, a governance to rebuild call for neither the same sources, nor the same agents, nor the same controls.

Built on Claude

Anthropic's frontier model, specialised in a single subject matter — leadership, governance, strategy, organisation. Every document read, every cross-check made, the watch kept continuously.

Governed

Every claim traces back to its document, its date and its formula. One agent is tasked with attacking the previous agent's conclusion. And the machine never assesses a person — that is contractual.

The machine reads everything — that is its job. It concludes nothing — that is ours.

02 / The matrices — the precision

Knowing what to look for, and where to look.

Power without a grid produces volume. Goban is the repertoire of our conceptual matrices — proprietary, sharpened engagement after engagement: they say what to look for in the mass, how to read it, and what to weigh it against.

A matrix is not a template for answers: it is a discipline of the question. For each cell it states what the chain must bring back, what counts as proof, and what that cell does not allow you to conclude — that is how the engine is governed, and why two engagements can be compared.

The first — human capital assessment

Six dimensions — Vision at the centre, four quadrants, Governance as the outer ring — and two loops linking the official company to the real one. The reading does not hand out grades: it locates the fulcrum, the point where a limited action reverses the dynamic.

A team is not the sum of its profiles — it is a system; the Goban is how it is read.

The next ones — in the writing

Selling, rebuilding a governance, halting a trajectory, committing a name, weighing a candidacy, preparing a succession, integrating after an acquisition: each family of decision calls for its own matrix — the documents to demand, the questions that determine, what counts as proof, and where the grid stops. They are written engagement after engagement; that is how the repertoire grows.

All of it anchored in what science validates: the structured interview, the Big Five family — the only instruments with documented validity — and the cybernetics of organisations.

One cell, unfolded in full: See “Critical Functions” →

03 / The human touch — the senior partners

What the machine will never do: commit.

A senior partner carries your file end to end — framing the question, conducting the interviews, contesting what the chain reports, settling what it leaves open, and signing. No juniors: the chain carries the volume, the partner carries the judgment. A committee of senior advisors, experienced executives and investors, brings a second eye to every reading.

The red line, contractual: the machine never assesses a person · no decision is automated · consent first, individual debrief after · every conclusion is signed

04 / Kifu — the decision cockpit

Every engagement has its cockpit.

An equation is only worth what it returns. What we return is not an end-of-engagement document: it is an open file, kept current, that you open from a link the day the question arises.

All of it is steered from Kifu, the cockpit of your decision: the state of the file, the deadlines counting down, the deliverables as they land, the log of what has changed — behind one link, with each reader opening exactly what concerns them.

Next step

The equation is judged on your situation.