Skip to content

Resources

Beyond the technical governance framework and implementation guide, we have also developed additional tools and resources to help organisations understand and get started with using the ARC framework to govern their agentic AI systems.

Resource Target Audience Description
ARCvisor Tool AI governance teams, developers Web-based risk assessment workflow with 50%+ time savings
Technical Paper AI governance teams, academia In-depth technical details and research foundations
Baseline Risk Register AI governance teams, developers Interactive exploration of 46 risks and 88 controls (YAML sources)
Presentation slides (pending) General AI-literate audience High-level overview of the ARC framework

ARCvisor 🤖✨

ARCvisor is an AI-powered assistant that makes risk assessment for agentic AI systems actually enjoyable. By combining LLMs with structured knowledge representation, ARCvisor turns the tedious process of risk identification and control selection into an interactive conversation.

Try Live Demo View on GitHub

🚀 Key Features

  • Automated risk identification: Identifies relevant risks based on system descriptions
  • Control recommendations: Suggests appropriate controls for specific deployment contexts
  • Documentation generation: Produces structured risk mitigation documentation
  • Conversational interface: Natural language interaction for describing agentic systems
  • Improved Efficiency: Reduces time required for manual risk assessment workflows

📚 Learn More

Want to dive deeper into how ARCvisor works? Check out the ARCvisor preprint paper for technical details on the architecture, evaluation results, and real-world case studies.