About Orchestrate

---

Open Governance for AI

Communities have always governed the systems that affect them, through laws, norms, and the right to change the rules when they cause harm. AI deployment broke that feedback loop. Governance gets baked into models at training time, locked in by vendors, and put beyond the reach of the communities subject to it.

Orchestrate offers communities a way to write enforceable governance standards for AI systems operating in their world. The burden of compliance sits with the companies deploying the AI, not with the communities. 

We do this by positioning an Independent Governance Layer between an application and the AI model beneath it, decoupling governance from the model itself. It is designed so that communities own it and vendors cannot override it. 

We offer stakeholders affected by a specific AI deployment (borrowers, regulators, advocates, practitioners) a framework for the co-authorship of a Community Constitution: a human-readable, legally grounded governance document that translates local social norms and regional regulations into measurable, machine-executable rules.

The Community Constitution is not a policy statement; it is a technical specification. It defines the AI's operating envelope through measurable rules, quantifiable escalation thresholds, appeal rights, prohibited data sources, and a formal amendment process. Each Constitution is written to be enforced upon the AI system itself, providing a clear standard for the companies that deploy it and the regulators and communities that hold them accountable.

AI providers and application developers must demonstrate that their systems operate within these guidelines, continuously, with auditable evidence for the stakeholders to assess. We propose a machine-verifiable handshake as the technical mechanism for that demonstration. Read more about the technical architecture of the Governance Layer. Using this framework, we offer a fundamental shift away from catch-up governance to embedded governance.

What we are not

We are not an AI vendor. We do not build the compliance infrastructure inside companies’ systems. That is their responsibility.

In a proof of concept using synthetic data, we found it technically feasible for companies to operate transparently and adhere to community standards without exposing their proprietary logic.

See it in practice 

We have drafted four illustrative constitutions covering credit assessment in Kenya, employment screening in Canada, work permit processing in the UAE, and healthcare diagnostics in Japan. Each is grounded in AI systems currently operating in that domain.

They are the starting point. The next step is for regulators, government AI oversight bodies, and affected communities to convene the stakeholders needed to negotiate, amend, and adopt the actual versions that will govern real systems. Read: Four Countries, Four Real Use Cases. 

In our Orchestrate Protocol post, we show a proof of concept using synthetic data for how communities can manage high-consequence automated decision systems through specific Constitutions. It shows how a Constitution can be integrated into the AI stack without requiring updates to the underlying AI.

This work is already reaching decision-makers. We have filed joint submissions with Kenyan legal and policy partners on the country’s AI Bill and its Draft AI Policy, submitted comments on the Central Bank of Kenya’s Draft Guidance Note on AI in the Banking Sector, provided written input to the UN Global Dialogue on AI Governance, and published a public response to Microsoft’s AI governance consultation.

Founders

Todd Simpson has built technology companies, led innovation at Mozilla and AVG, invested in emerging technologies at Inovia Capital, and currently serves on the board of the Alberta Machine Intelligence Institute.

Anar Simpson is a computer scientist who designed PanCanadian Petroleum’s first AI expert system, then led digital inclusion work worldwide. As Technovation’s Global Ambassador she expanded the program to 100+ countries. She has served as a Deputy to the UN Secretary-General’s High-Level Panel on Women’s Economic Empowerment and as an Advisor to Mozilla and the U.S. State Department’s TechWomen program.