A process model says who does what, in what order, and how success is tested. It is the operating layer between a governance framework and a piece of software. The framework sets the outcomes. The process model turns those outcomes into work. The software runs the work and keeps the record. Without the middle layer, software implements activity with no test of success.
What a process contains
A process in this sense has a control objective, a trigger, an owner, inputs, activities and a work product. The control objective says what has to be true. The trigger says when the work starts. The owner is accountable for the step. The inputs are what the step needs. The work product is the record the step leaves. A measure says how the organisation knows. Assurance is the sample of that work.
Order matters. Intake opens the work. Context, policy and design come before build. Build comes before operation. Operation produces the monitoring and the incidents that review samples. A material finding returns to direction and can open new work. That line is how a regulation written as a system lands without a parallel factory for each duty.
Depth is not uniform. Design factors set how much of each process is implemented, and to what depth. A low-risk internal tool does not take the same process as a system that changes a customer outcome or processes personal information under POPIA. The objectives stay the same. The practices scale. A model that implements every process at the highest intensity for every use is theatre. A model that concentrates depth where risk and value are highest is a system.
DAEIOR
DAEIOR is the process model. The name is the six domains: Direct, Act, Establish, Implement, Operate and Review. It is the process library behind Control Objectives for AI Systems (AICOB). AICOB is how the organisation establishes, implements, operates, monitors, reviews, maintains and continually improves its AI ecosystem. DAEIOR is the ordered set of processes that does that work.
Direct is the governing body. It evaluates, directs and monitors, and holds management to account under King V. Act is management intake: a business need, a board direction, an incident, a vendor event, drift, or a change of intended purpose. Establish sets context, policy, objectives, people, vendors, risk, security, data and the design of the system. Implement designs, acquires, builds, tests, changes and deploys. Operate runs the system, handles incidents and continuity, and attaches policy to live systems. Review samples performance, conformance, internal control and assurance, and returns a material finding to Direct.
Governance and management stay apart. Direct does not build the management system. Establish does not set appetite. Review does not write the charter. An agent is never accountable. Residual-risk acceptance stays with the governing body. A notice is not a control.
DAEIOR is intellectual property. It is embedded in the software. The intimate knowledge of the process is the consulting. The library itself is not published as a catalogue. The public statement is the structure: six domains, named processes, named owners, named work products.
What the model gives
One accountability model. Every material outcome has a governing-body role, to evaluate, direct and monitor, and a management owner who runs the step.
Business language for technology. Goals cascade from enterprise objectives, so a control can be shown to exist for a reason, not only that it exists.
A complete lifecycle, not a control list. Strategy, architecture, build, run, continuity and assurance sit in one model, together with governance, portfolio, people, vendors and change.
Auditability. Practices produce named work products. Internal audit, the Information Regulator and the governing body can sample a process, not a slide deck.
Reuse across obligations. One process-control activity can support more than one instrument when the mapping is explicit. DAEIOR is the spine other instruments map onto, including King V, POPIA, and the EU AI Act where it reaches the organisation, as well as ISO/IEC 42001. A mapping is a claim only when it is explicit: one process named against a named article, clause or control.
A stable spine when technology changes. Agents and models change quickly. Benefits, risk, change, incidents and performance do not.
The platform
The platforms that run this model are process-based. The product name changes with the topic. The method does not. The EN 18286 quality management system (where the EU AI Act applies), the ISO/IEC 42001 AI management system, and the systems that support King V and POPIA all run on DAEIOR. The platform holds what is running and whether the work was done. Software may automate inventory, policy, runtime guardrails, monitoring and evidence. It does not replace the governing body's duty to evaluate, direct and monitor.
None of this certifies the organisation or declares it compliant. Those decisions sit with the assessor, the Information Regulator, a notified body where the EU AI Act applies, or the governing body itself. The system makes the work visible so the decision can be made on evidence.