ACO
ACO is a governance and orchestration layer for multi-agent work. It is designed to package approved context, route it to connected AI tools, and return outcomes with provenance and execution evidence. Connector availability and capabilities vary by integration; production use is subject to technical validation and governance requirements.
Let's start a great collaboration
Agentic orchestration
ACO is designed to assemble only the context approved for a task and deliver it to a selected agent or connected tool. This supports a clearer boundary between source material, instructions and the work a connector is allowed to perform.
Control model
A structured control layer for coordinating context, tools and review points across agentic workflows.
Maps an approved context package to a defined workflow or connector.
Keeps references to source material and transformation steps available for review.
Expresses operational constraints that can be checked before and during execution.
Supports review gates for decisions that require accountable human approval.
Explicit boundaries
ACO is designed to make orchestration boundaries explicit and to support deployment patterns aligned with an organisation's control requirements. It complements—not replaces—enterprise identity, security and compliance controls.
Traceable exchange
Connected tools can return outputs, status and evidence into a shared execution record, enabling teams to inspect what was requested and what came back. Available evidence depends on the verified capabilities of each connector.
Controlled adoption
ACO is being prepared for controlled integrations and pilot validation. Supported connectors, operating boundaries and service levels should be confirmed during technical and governance acceptance before production use.