The Organizational Plane
Why agent builders and control towers do not fully answer who owns the outcome.
Research and Category Development
From agent systems to accountable computational institutions.
Coda’s research explores how mandates, roles, decision rights, authority, evidence and performance turn autonomous capabilities into durable organizational capacity.
Read →
The Accountable Guild
A portable institution combining mandate, roles, sponsorship, authority and evidence.
Verified Closure
Why alerts, agent runs and completed actions are insufficient economic units.
Executable assurance standards
Translate safety and governance requirements into architecture, policy, roles, tests and evidence.
Do not deploy a swarm. Design an institution.
Six principles for durable human–AI operating systems.
Computational institution manifesto
Six principles that remain foundational.
Every capability needs a sponsor.
An agent may perform work. A named human or institution remains accountable.
Roles survive model replacement.
The organizational role remains stable as models and vendors change.
Decision rights must be explicit.
Proposing, verifying, approving, executing and vetoing are distinct.
Human authority is designed.
Human involvement is an operating model, not a generic review node.
Reputation comes from evidence.
Performance derives from verified outcomes, not self-reported claims.
Institutions retain memory.
Decisions, exceptions and learning survive individual agents.