AI-related regulations and standards implementation and assurance

Overview

Each AI-related regulations and standards implementation and assessment workbench directs an EN/ETSI/ISO/NIST standard or DOR/NIS2/CRA regulation implementation and assessment across the organisation, using a process-based approach. Process owners work from action lists, determine requirements, risks, controls, implementation roadmap, artefacts and compliance.

A process flow approach is adopted for the implementation, for example:

• Cyber Resilience Act
  • Scope
  • Requirements
  • Vulnerabilities
  • Conformity
  • Information sharing
• DORA
  • ICT risk
  • Incidents
  • Testing
  • Third parties
  • Information sharing
• NIS2
• EN standards
• ETSI standards
• ISO standards
• System development life-cycle best practice
• Agent-control best practices
  • Direct the harness
  • Plan the agent
  • Code and build
  • Before the agent acts
  • After the agent writes
  • Test, certify and ship
  • Deploy and operate
  • Monitor, handle incidents and retire
• Multi-agent system best practices
  • Direct the multi-agent harness
  • Plan the graph
  • Identity, non-human identity and delegation
  • Orchestrate and message
  • Shared tools, MCP and memory
  • Before and after hops
  • Certify the graph and ship
  • Operate the mesh
  • Monitor, handle incidents and retire