What is the difference between an AI CoE and a Power Platform CoE?

An AI Centre of Excellence governs decisions about AI: which use cases proceed, at what risk tier, with what human oversight and what evidence. A Power Platform Centre of Excellence governs the platform those things are built on: environments, data loss prevention, application lifecycle, licensing and maker enablement. Large organisations usually need both, and the platform CoE normally comes first because the AI council depends on its inventory and guardrails.

Do we need an AI CoE if we already have a Power Platform CoE?

If Copilot Studio agents or custom models are reaching production, yes. A platform CoE tells you what exists and keeps it inside a boundary. It does not decide whether a given AI use case is acceptable, who signed it off, or how it is evaluated before release. Those are the questions risk and legal ask, and they need a council with named decision rights.

How long does it take to stand up?

The council charter, intake process and risk tiering can be agreed and running within the first quarter. The register of existing AI assets depends on how much is already in the tenant. We normally run the first quarter with you and then hand the operating rhythm to your team.

Which standards does the assurance model follow?

Our responsible AI policy is mapped to ISO/IEC 42001 and Australia's AI Ethics Principles. Mapped, not certified: the CoE produces evidence that an auditor will accept, and it does not by itself certify your organisation against a standard.

AI Centre of Excellence

Run AI like a portfolio, not a pile of pilots.

An AI Centre of Excellence for organisations past the pilot stage. We stand up the operating model, the assurance controls and the platform guardrails that let thousands of people use Copilot, agents and Power Platform without IT losing sight of any of it.

What an AI Centre of Excellence governs

Most large organisations now run dozens of AI pilots and none of them at scale. Copilot licences sit unused. Agents appear in Copilot Studio built by whoever had the idea that week. Nobody can name every model touching customer data, so risk and legal ask who signed this off, and the honest answer is nobody did. Then the board asks what the AI spend returned this year.

We stand up the CoE that answers those questions. That means an AI council with named decision rights, a use-case pipeline scored on value and risk, guardrails on the platforms your people already use, and one register of every AI asset in the tenant. We run it as a managed service for as long as you need, then hand it to your team. The CoE decides what gets built and under what controls; building the agents themselves is a separate engagement.

Operating model and AI council

Who decides, who funds, who carries the risk. We design the council and the intake process, then run the first quarter with you.

Council charter, membership and meeting cadence
Use-case intake scored on value and risk
Funding model and chargeback to business units
One register of every AI use case, agent and model in the tenant

Policy, risk and assurance

The controls that let legal and risk say yes. Each use case gets a risk tier, human oversight where it counts, and evidence an auditor will accept.

Responsible AI policy mapped to ISO/IEC 42001 and Australia's AI Ethics Principles
Risk tiering with matching approval gates
Evaluation and red-team testing before release
Audit evidence pack per use case

Platform guardrails

Environment, data protection and lifecycle controls across Microsoft 365 Copilot, Copilot Studio and Power Platform, so makers move fast inside a boundary.

Environment strategy for makers, agents and production
DLP and Purview policies tuned to your risk appetite
CoE Starter Kit inventory across apps, flows and agents
ALM pipelines with managed environments

Adoption, skills and value tracking

Capability grows or the platform stalls. We run the training pathway and the champions network, and report what each use case returned.

Role-based learning paths for staff, makers and agent builders
Champions network, office hours and accreditation
Copilot licence assignment based on measured usage
Benefit tracking per use case, reported to the council

What changes once the AI CoE is running

No shadow AI

Every agent, app and model in the tenant sits in one register with a named owner.

Faster approvals

Risk tiers and standing decision rights replace a case-by-case argument with legal.

Licence spend that works

Usage data shows which Copilot and Power Platform licences earn their keep, and which get reassigned.

Who an AI Centre of Excellence is for

Built for:

CIOs and Chief Data or AI Officers
Heads of Digital running an AI programme past its first pilot
Risk, legal and privacy teams asked to sign off on AI
Enterprise architects and platform owners
CFOs asking what the AI spend returned

Runs on

Microsoft 365 Copilot
Microsoft Copilot Studio
Microsoft Foundry
Microsoft Power Platform
CoE Starter Kit
Azure DevOps / GitHub for ALM

Built to align with

Designed for IRAP PROTECTED environments
Australian Privacy Principles
Essential Eight aligned
ISO/IEC 42001 aligned

Accelerate Your Desired Outcomes

Clients choose us because we:

Are certified Microsoft experts.
Take a holistic approach to understand your business needs and bridge the gaps with the right tools and solutions.
Collaborate with you and prioritise your input to ensure a client-centric approach.
Operate with agility and speed, thanks to our small teams and 2-week sprint model.
Start developing on day 1 to get results quickly.
Focus on delivering outcomes, not just scope, to ensure we meet your goals.
Offer cost-effective solutions that enable you to reach your goals faster.
Stay flexible and open to new and better ways of doing things to continually improve your business processes.
Are proudly Australian owned and operated.

Frequently asked questions

What is the difference between an AI CoE and a Power Platform CoE?: An AI Centre of Excellence governs decisions about AI: which use cases proceed, at what risk tier, with what human oversight and what evidence. A Power Platform Centre of Excellence governs the platform those things are built on: environments, data loss prevention, application lifecycle, licensing and maker enablement. Large organisations usually need both, and the platform CoE normally comes first because the AI council depends on its inventory and guardrails.
Do we need an AI CoE if we already have a Power Platform CoE?: If Copilot Studio agents or custom models are reaching production, yes. A platform CoE tells you what exists and keeps it inside a boundary. It does not decide whether a given AI use case is acceptable, who signed it off, or how it is evaluated before release. Those are the questions risk and legal ask, and they need a council with named decision rights.
How long does it take to stand up?: The council charter, intake process and risk tiering can be agreed and running within the first quarter. The register of existing AI assets depends on how much is already in the tenant. We normally run the first quarter with you and then hand the operating rhythm to your team.
Which standards does the assurance model follow?: Our responsible AI policy is mapped to ISO/IEC 42001 and Australia's AI Ethics Principles. Mapped, not certified: the CoE produces evidence that an auditor will accept, and it does not by itself certify your organisation against a standard.