Humans stay accountable
AI prepares, summarises and drafts. People decide. Every system has named owners and explicit review points.
A method you can hold us to
Every CedarPro engagement runs through the same seven stages. No black boxes, no mystery phases — you always know where the work is and what evidence the next stage needs.
The method
Seven stages, one discipline
1
Assess
We start by understanding how work actually happens — not how the org chart says it does. Discovery workshops surface the workflows, frustrations and decision points; the AI readiness assessment establishes capability, data and governance baselines.
- Discovery workshops
- AI readiness assessment
- Workflow mapping
- Capability baseline by team
2
Design
Before anything gets built, we agree what good looks like: the operating model, the governance and acceptable-use rules, and a roadmap with pilots ranked by value and feasibility.
- AI operating model design
- Governance and risk controls
- Prioritised roadmap
- Success measures agreed up front
3
Build
Agents and AI systems are built or configured using a co-build approach wherever possible — your people in the room, learning as the system takes shape. Prototypes are tested against real documents and real cases.
- Co-build workshops
- Prototype with real users
- Permissions and security by design
- Iterate prompts, tools and outputs
4
Train
Capability transfer is a delivery stage, not a courtesy. The people who will use, improve and govern the system are trained on it directly — at their level, in their roles.
- Role-based training
- Verification and quality habits
- Internal champions identified
- Playbooks and usage guides
5
Implement
The system goes into live workflows with support on hand. Adoption routines — stand-ups, office hours, feedback loops — carry teams through the awkward early weeks.
- Staged roll-out
- Adoption support and routines
- Issue triage and refinement
- Handover playbook
6
Govern
Governance keeps the system trustworthy: who can use it for what, where human review is mandatory, what gets logged and who is accountable. Public sector clients get audit-ready documentation as standard.
- Acceptable-use rules
- Human-in-the-loop checkpoints
- Audit trails and logging
- Clear accountability
7
Improve
Measurement closes the loop. We track usage, time saved and workflow outcomes against the measures agreed at design — and refine, retire or extend based on the evidence.
- Benefit and adoption tracking
- Reporting for leaders
- Continuous refinement
- Next-wave roadmap
Principles
What stays true in every engagement
Governance is not optional
Permissions, audit trails and acceptable use are designed in from day one — never bolted on after an incident.
Capability transfer over dependency
We measure success partly by how little you need us at the end. Your team should be able to run, improve and govern what we build together.
Questions
About the method
Usually two to four weeks for a single service area: discovery workshops, a readiness assessment and workflow mapping. Larger organisations run it per department rather than stretching one assessment across everything.
Yes — most clients do. A readiness assessment, a one-day intensive or a single agent co-build are all self-contained starting points. Each produces something useful on its own and tells you whether a bigger engagement is justified.
Always. IT, information governance and legal are involved from the Design stage, not presented with a finished system. In the public sector that includes producing the documentation your governance processes require.
See the method applied to your workflows
A discovery call is stage zero. Bring one stubborn workflow and we will walk you through how the seven stages would handle it.