Separation of powers is the whole point. The people who propose strategies are structurally prevented from being the people who approve them, and permissibility sits above everything in a reserved authority.
Hover or tap a role to see what it owns. Each is a distinct seat of authority, and the boundaries between them are defended, not blurred.
ASQIOS uses an AAOIFI-primary methodology. The Shariah Supervisory Authority holds a reserved veto over every permissibility question, and any new strategy mechanism requires a fresh ruling from qualified human scholars before it can be researched.
Compliance is the first priority in the decision order, not a screen applied at the end. A strategy that is not permissible never reaches the question of whether it performs.
The compliance layer is not something the research team can argue its way around. A ruling is required input, the screen is not re-performed by downstream AI, and the authority to decide permissibility sits in one place by design.
The platform is governed by a documented constitutional corpus — frameworks for risk, validation, capital allocation, strategy lifecycle, AI governance, Shariah compliance, data, and execution. Amendments move through a defined, recorded cycle rather than informal drift.
The top-level instrument every other framework answers to, kept internally consistent across revisions.
Each domain has its own framework with explicit failure modes named without euphemism.
Revisions are proposed, reviewed, and closed on the record — so governance is auditable, not ambient.
Where one person holds several roles, the separation-of-powers caveat applies throughout: no seat approves its own output.