The problems
we're built for.

Most consulting firms organize around service lines. Epirroi organizes around problems. The distinction matters because the problems that reach this practice do not fit neatly into a single discipline. A sovereign AI strategy engagement requires behavioral science, institutional design, technology architecture, and geopolitical navigation — simultaneously. A defense technology commercialization mandate requires understanding DoD procurement logic, GCC trust sequencing, market positioning, and regulatory pathway design — in the same room, not across four separate workstreams billed independently.

The mandate model means every engagement is scoped around the actual problem the institution is facing, not around the capabilities Epirroi happens to sell. If the diagnostic reveals that the bottleneck is behavioral — the institution knows what to do but cannot make itself do it — the engagement focuses there. If the bottleneck is architectural — the decision systems don't exist — the engagement builds them. If the bottleneck is market-facing — the technology works but the go-to-market sequencing is wrong — the engagement fixes the sequence.

This is also why every engagement begins with a paid diagnostic. The diagnostic is not a sales exercise. It is a standalone deliverable that maps the institution's strategic friction, decision architecture, AI readiness, and 90-day horizon. The client keeps the diagnostic regardless of whether they proceed. If the problem is not a good fit for Epirroi, the diagnostic says so. This model ensures that the engagement that follows is built on institutional truth, not on a proposal designed to win business.

The two-week diagnostic sprint is the same regardless of mandate type. It produces four deliverables. First, an institutional architecture map — how decisions actually get made, where authority sits, what the formal and informal power structures are, and where the gaps between stated strategy and actual behavior exist. Second, an AI readiness assessment — not a technology audit, but an evaluation of whether the institution's decision systems, data governance, and human workflows can actually absorb AI tools without creating new problems. Third, a behavioral barrier analysis — identifying the specific beliefs, incentive structures, and institutional habits that are preventing the strategy from executing. Fourth, a 90-day action sequence — not a roadmap, but a prioritized sequence of moves that accounts for political constraints, resource availability, and institutional willingness to change.

The diagnostic is designed to be useful even if the institution never engages Epirroi again. It is a decision tool, not a pitch deck. Clients have used diagnostics to restructure internal teams, redirect technology investments, reframe board-level strategy, and — in several cases — to decide not to pursue the initiative at all, which saved them significantly more than the diagnostic cost.

After the diagnostic, engagements typically run 3–12 months depending on mandate complexity. Sovereign AI strategy mandates tend toward 6–12 months because institutional change in government moves at a different clock speed. Defense technology commercialization mandates can be shorter — 3–6 months — when the technology is ready and the market entry sequence is the bottleneck. Crisis-driven transformation mandates are the most compressed, sometimes requiring 30–60 day execution windows where the cost of delay is measured in institutional relevance, not revenue.

Every engagement is principal-led. Michael Joseph is in the room for every critical decision point, not managing from behind a staffing model. For mandates requiring specialized depth — agentic AI architecture, regulatory pathway design, GCC market access — Epirroi brings in domain specialists from an established network of operators who have built in the field. The team scales to the mandate. The principal does not disappear after the sale.