Data and operations
Stages
- 03 Prioritise
- 05 Operationalise
- Problem
- The data we need is scattered, without owners or reliable quality.
- Result
- Data sources with owners and quality gaps identified.
Stage 5 of 6
Operationalise
Operating model
The question
How do we build governance into operations?
You arrived from the home page with this situation: “We have rules, but they aren't applied”. Choose another situation
The problem it solves
Objective Build the rules into everyday work.
The policy exists in a document but has not changed how things are bought, built or approved. Roles are defined on paper and each area handles it its own way.
Questions it answers
- How do we build governance into operations?
- How do we bring governance into existing processes?
- Who decides, and what gets recorded, when a new AI use is approved?
Signs your organisation is at this stage
- We have approved rules, but they haven't changed how we buy, build or approve.
- Roles are defined on paper and each area handles it its own way.
What AGORA does
- Surveys how AI uses are approved today, who decides and what gets recorded.
- Designs the decision model and approval workflows.
- Builds control points into the procurement and development cycle.
- Accompanies the roll-out and verifies the evidence the controls produce.
Integration points
- Buying tools.
- Vendor assessment.
- System development and changes.
- Approving new use cases.
- Data handling.
- Incident management.
- Reporting to leadership.
Deliverables
- A decision model and approval workflows for new AI uses.
- Control points embedded in the procurement and development cycle, each with an owner.
- Defined evidence produced by each control and where it is recorded.
- A review cadence and reporting to leadership.
Who needs to take part
- The areas that buy, build and approve AI uses.
- Whoever takes on the roles in the decision model.
The organisation assigns the roles and adapts its processes. AGORA designs the model, accompanies the roll-out and verifies that controls produce the agreed evidence.
Indicative duration
Defined by scope, based on what the assessment recommends.
Exit criterion
- Approval workflows defined for new AI uses.
- Controls embedded in the procurement and development cycle, each with an owner.
- Evidence recorded at every step and periodic reporting to leadership.
Solutions involved
Next step
Not a compulsory process
You do not have to go through the six stages in order, or engage work at every stage. An organisation with systems already in production can come in at Govern or Operationalise; one that is just exploring usually starts by understanding and assessing.