rAI: Governed AI Task Forces for Real-World Operations

Project summary

rAI (Responsible AI) is an open governance framework for organizing autonomous AI agents into trustworthy, accountable task forces that can safely coordinate real-world operations. While AI capabilities continue to advance, there is comparatively little practical work on how autonomous agents should be governed as participants in organizations and markets. This is where rAI comes in.

Nova, the initial reference implementation of rAI for commerce, provides a real-world environment for validating and refining the framework through production deployments.

What are this project's goals? How will you achieve them?

This project develops the governance models, reference implementations, and open-source infrastructure needed for reliable multi-agent coordination.

At the core, is a structured multi-agent framework where specialised AI agents operate through defined responsibilities, governed collaboration and accountable decision making.

The vision is to create an open coordination protocol for governed multi-agent systems.

The rationale: the more capable AI becomes, the greater the consequences of mistakes.

rAI would therefore ensure more capable autonomous AI systems can safely accomplish real-world work.

Who is on your team? What's your track record on similar projects?

I am currently the sole full-time researcher and developer behind rAI and the founder of Univus Cloud. Over the past two years, I have been developing the concepts behind rAI while designing and building Nova, its initial reference implementation for commerce, to validate the framework through practical deployments.

My background is in software engineering, distributed systems, cloud infrastructure, and AI-powered applications. Throughout my career, I have focused on building systems that reduce coordination costs and improve operational efficiency for organizations. During the past eight months, I have worked full time on Univus Cloud, developing a functional AI-native commerce platform, securing early commercial engagement, and continuing to refine the governance concepts that underpin rAI through Nova.

What are the most likely causes and outcomes if this project fails?

The primary risk is that the governance abstractions I am developing may not generalize as effectively as anticipated across different domains or scales of deployment. Real-world testing may reveal that certain governance models, coordination patterns, or oversight mechanisms require significant redesign before they become broadly applicable. Adoption may also be slower than expected if organizations perceive governance as introducing additional complexity.

Even if these challenges arise, I expect the project to generate valuable public goods. The open-source framework, governance specifications, technical documentation, deployment case studies, and lessons learned from production environments will contribute practical knowledge to the growing field of multi-agent AI governance. Demonstrating which approaches do not work is itself a meaningful research outcome that can inform future work by both researchers and practitioners.

How much money have you raised in the last 12 months, and from where?

$0

Funding milestones

US$12,000 (Minimum Funding)

Supports approximately three months of full-time research and development. This milestone will deliver an expanded core governance framework for rAI, an initial open-source release, governance documentation, and early production validation through Nova, the initial reference implementation.

The microgrant will fund the development and validation of the rAI framework via AI model credits, cloud infrastructure and dedicated engineering/research time.

Outputs will include: a reference implementation, documentation and implementation guides.