Keeping twelve client strategies alive between two missions

Every hoshin practitioner knows the curve: workshop enthusiasm in January, a beautiful matrix on the wall, and by the third month the spreadsheet is stale because nobody on the client side re-types execution data into it. The mission ends, the ritual decays, and next year's audit finds the matrix exactly where you left it.

What survives is not the diagram — it is the cadence plus a tool that keeps the data fresh without depending on anyone's discipline. That is the gap this platform closes: the matrix stays linked to the client's execution tools, and an agent does the weekly maintenance.

Each client is a separate organization: their own matrices, users, roles and audit log, strictly isolated from every other client (multi-tenant isolation is enforced on every query).

How you move between them, precisely, because it is a real constraint and you should hear it from us rather than discover it: an account belongs to one organization. There is no single login that switches between client workspaces today. The practical pattern is one address per engagement — `you+acme@yourfirm.com`, `you+globex@yourfirm.com`. Mail providers deliver all of them to your inbox, each is a distinct member of a distinct client organization, and each gets its own API token.

That matters less than it sounds, and the reason is the whole thesis of this product: your agent holds those tokens, so it can read Acme's matrix and Globex's matrix in the same conversation, score both and prepare both agendas. The thing you cannot do is have one browser session that switches workspaces — and if the agent does the maintenance, you open the web interface a few times a month. In sovereign local mode the question disappears entirely: one data directory per engagement, no accounts at all.

One caveat worth stating: reusing the *same* address across two organizations is not the pattern to adopt. Sign-in resolves the address to a workspace, and a duplicated address is refused rather than resolved arbitrarily — distinct aliases keep every login unambiguous.

Light white-label puts the client's logo on their workspace and exports. And the honest clause consultants rarely get from software vendors: when a client matures, they graduate to their own subscription and keep all their data. Your value is the method and the coaching — not a hostage-taking license.

One consultant, eight clients, eight weekly reviews to prepare — that arithmetic is why hoshin missions stay small. An AI agent connected through MCP changes it. For each client organization, the built-in weekly_sync prompt:

• Reads the client's initiatives and their linked execution projects (Linear, Jira — via each tool's own MCP).
• Updates initiative statuses and flags drift: execution work that maps to no initiative, initiatives linked to nothing.
• Drafts the review agenda: off-track KPIs, blocked initiatives, decisions pending since the last session.
• On demand, runs a deployment diagnosis on any strategic theme — how deep does it actually cascade?

The completeness score computes five metrics on any matrix: quadrant coverage, correlation density, orphan elements, vision-to-KPI chains, and leading/lagging indicator balance. For a coach this is gold: the critique is objective and non-personal. You do not tell the COMEX their strategy is incoherent — the score does, with the exact list of orphans and broken chains.

It also makes your impact measurable across sessions: the score's trajectory from the first workshop to the locked matrix is a before/after you can put in a case study.

The Partner tier is €290/month with three client organizations included, then €49/month per additional client — unlimited matrices per client, assisted onboarding, light white-label. During the beta everything is unlocked, and founding partners keep an advantage when billing starts and get a real say in the roadmap: the features consultants need (client templates, cross-client dashboards) are prioritized by the cabinets actually using it.