Scenario: Launch a CoE¶
A CoE without a governance model is a team. A governance model without a CoE is a document. You need both.
TL;DR
Launching a CoE follows a specific sequence: establish ownership first (SCALE-OPS Stewardship), define the environment structure (SCALE-OPS Containment), set the delivery model (BOLT), establish data and connector guardrails (SHIELD Harden), then build operational capability (SCALE-OPS Operations, Automation). Skipping the sequence creates gaps that are expensive to close later.
Applies To
Audience: CoE Lead · Platform Admin · CIO · IT Director Situation: Starting a new CoE, or formalising an informal one that has grown without structure Frameworks: SCALE-OPS · BOLT · SHIELD
The Situation¶
Your organisation has adopted Power Platform — possibly organically, possibly through a deliberate initiative. There are makers, there are solutions, and there is growing pressure to govern what has been built and enable what comes next. You have been asked (or have decided) to set up a Centre of Excellence.
The CoE is not a team. It is a governance capability — the combination of ownership, tooling, process, and community that enables the platform to scale safely. Getting the sequence right matters: a CoE built in the wrong order will have structural gaps that become crises later.
The Sequence¶
Step 1 — Establish ownership (SCALE-OPS: Stewardship)¶
Before anything else, clarify who owns what. This is SCALE-OPS Stewardship — the people and structure layer of platform governance.
Key decisions at this step:
- Who is the Platform Owner? (executive accountability)
- Who runs the CoE operationally? (CoE Lead)
- How does the CoE relate to existing IT governance?
- Which business units have a Fusion Team Lead?
Don't skip this. A CoE without clear ownership defaults to the person who shouts loudest, which means governance is reactive rather than deliberate.
Step 2 — Define the environment structure (SCALE-OPS: Containment)¶
Environment strategy is the single most commonly misconfigured aspect of Power Platform governance. Set it early — before makers have built in the wrong places.
Key decisions at this step:
- How many environments? What types? (Dev / Test / Prod minimum)
- Which environments are Managed?
- What is the default environment policy?
- Where do Tier 1 makers build?
Common mistake: Starting with one environment and trying to split later. Solutions proliferate in the default environment and cannot be easily migrated.
→ SCALE-OPS: Containment · Environment Tier Decision Guide
Step 3 — Set the delivery model (BOLT)¶
Once environments exist, define who builds what and how. This is BOLT — the operating model for business-IT collaboration.
Key decisions at this step:
- Which delivery tiers will you operate? (Tier 1 only, or all four?)
- What are the role definitions for Maker, Fusion Team Lead, Platform Owner?
- What guardrails apply at each tier?
- What is the pre-approval process for connectors and data?
→ BOLT: Delivery Tiers · BOLT: Roles & Accountability · BOLT: Guardrails
Step 4 — Establish data and connector guardrails (SHIELD: Harden)¶
Before makers build in earnest, define the data protection baseline. DLP policies, data classification, and connector governance must be in place before adoption scales — not after.
Key decisions at this step:
- What connectors are approved for each environment tier?
- What is the DLP policy structure?
- Which data classifications exist? What controls apply to each?
→ SHIELD: Harden · DLP Policy Decision Guide
Step 5 — Build operational capability (SCALE-OPS: Operations, Automation)¶
With structure in place, build the operational rhythm: monitoring, health reviews, automated compliance checks, and CoE toolkit deployment.
Key decisions at this step:
- Which CoE Starter Kit components will you deploy?
- What is the cadence for platform health reviews?
- How will compliance drift be detected and resolved?
- What is the incident response process?
→ SCALE-OPS: Automation · SCALE-OPS: Operations
Step 6 — Enable and grow (SCALE-OPS: Enablement)¶
Once governance is operational, invest in the maker community. Training, office hours, communities of practice, and feedback loops accelerate safe adoption.
Common Mistakes¶
| Mistake | Consequence | How to Avoid |
|---|---|---|
| Starting with training before governance | Makers build in ungoverned environments | Establish environment structure first |
| One DLP policy for all environments | Dev is over-restricted; prod is under-protected | Tier DLP policies to match environment risk |
| CoE Lead with no executive backing | Cannot enforce decisions | Establish Platform Owner accountability first |
| Deploying CoE toolkit before defining metrics | Toolkit generates data nobody acts on | Define operational KPIs before deployment |