💼 Monday Business

GPM P5™ in Practice — Tracking Sustainability on a Live AI Program

June 15, 2026

I set a 2026 goal to apply GPM P5™ to live IT projects. Here’s what that looks like in practice — the kickoff, the metrics framework I built, and the first five months of results from an enterprise-wide AI enablement program.

From Goal to Real Project

Earlier this year I documented my 2026 professional development goals — one of which was to integrate GPM P5™ sustainability metrics into live IT projects. Not in theory. On real projects, with real data, reported publicly.

This is the first update.

The project is an enterprise AI enablement program: rolling out Microsoft Copilot across the organization, establishing data loss prevention (DLP), defining a Power Platform governance strategy, and building the AI governance framework that makes all of it safe to use at scale.

It’s a program, not a single project. Workstreams keep being added as the scope of “enabling AI safely” becomes clearer. That’s actually a more honest test of whether P5 metrics are practical — if they only work on clean, well-scoped projects, they’re not very useful.

Why AI Enablement Is a Sustainability Project

This isn’t obvious at first. AI enablement sounds like a technology rollout, not a sustainability initiative. But look at it through the P5 lens and the picture changes quickly.

The problem we’re solving: Copilot is smart enough to search across an organization’s data — but that’s only safe if employees only have access to data they’re supposed to have. In most organizations, permissions have drifted over years. People have access to things they don’t know they have access to. Turning Copilot loose on that environment doesn’t just create a security risk — it creates a data governance problem that could expose sensitive information at scale.

So before we can enable AI broadly, we have to get the foundation right: clean up data access, implement DLP, establish governance, and then expand Copilot’s reach as each layer is secured.

That sequencing — governance before capability — is a sustainability principle. It’s the difference between fast and reckless, and fast and durable.

The P5 Metrics Framework

When I kicked off sustainability tracking on this program, I built a metrics table anchored to P5’s five dimensions. Baseline was established at program start. Targets were set based on what success looks like. Data points update every one to three months — not every week, because these are trend metrics, not sprint metrics.

Here’s what I’m tracking and why each metric was chosen:

People

Who is being affected and how well are they supported?

Copilot adoption is only meaningful if people are actually using it — and only sustainable if they’ve been trained to use it safely. Handing someone a powerful tool without training isn’t enablement, it’s a liability.

Metric Baseline Target Mar Update Change
Active Copilot Users 144 400 327 +127%
Licenses Assigned 154 500 362 +135%
Training Complete 32% 90% 98.1% +66 pts
Copilot Agents 199 TBD 267 +68

Training completion at 98.1% against a 90% target is the number I’m most proud of. The goal was never to just assign licenses — it was to assign licenses to people who know how to use them responsibly.

Planet

What is the environmental and resource impact?

For a software program, “planet” metrics aren’t about carbon directly — they’re about time and resource efficiency. Meeting hours saved is a proxy for one of the most wasted resources in any organization: people’s time.

Metric Baseline Target Mar Update Change
Meeting Hrs Saved/mo 31 TBD 297 +858%

31 hours saved per month at baseline. 297 hours saved per month five months later. That’s a nearly 10x increase. Source: Microsoft 365 Copilot Dashboard (Viva Insights).

A target of TBD here is intentional — I didn’t know what was possible at baseline. Now I do. The next update will include a target based on what we’ve learned.

Prosperity

Is this creating real organizational value?

Actions per month measures how much actual work is getting done through Copilot — not just whether it’s been assigned, but whether it’s being used to accomplish things.

Metric Baseline Target Mar Update Change
Total Actions/mo 13K TBD 28K +114%

Doubling actions per month in five months. Again, TBD on the target — the baseline told us where we were starting, and we’re now building the trajectory that will inform a realistic target.

Process

Are our operational processes improving?

DLP is the foundational process layer. Email DLP via Proofpoint is complete. File DLP via AvePoint is in progress. Power Platform governance is being built in parallel.

Metric Baseline Target Mar Update Change
Email DLP 0% 100% 100% Complete
File DLP 0% 100% 0% In Progress
Power Platform Cloud Flows 447 TBD 464 +17

File DLP is the current blocker for expanding Copilot’s reach into SharePoint. Until files are governed, Copilot search stays restricted. This is the right call — it’s the governance-before-capability principle in practice.

Products

What are we actually delivering?

Metric Baseline Target Mar Update Change
Copilot SharePoint Search Blocked Enabled Blocked Pending
Power Platform Apps 63 TBD 92 +29

SharePoint search remaining blocked is the honest part of this update. We haven’t delivered the flagship capability yet. That’s not failure — it’s sequencing. The DLP work has to land first. But it’s worth naming clearly: the most visible product milestone is still pending.

Power Platform apps growing from 63 to 92 shows that the platform is being adopted even while governance work continues.

What the First Five Months Taught Me

P5 metrics work on messy programs. I wasn’t sure they would. AI Enablement doesn’t have a clean end date. Workstreams keep getting added. The scope is evolving. And yet the P5 framework gave me a way to track whether we were moving in the right direction across all five dimensions simultaneously — not just on the workstreams that were easiest to measure.

Some targets have to come after baseline. I set several targets as TBD at kickoff because I genuinely didn’t know what was achievable. 297 meeting hours saved per month wasn’t a number I could have predicted from a baseline of 31. That’s fine — TBD targets become informed targets once you have trend data. The next reporting cycle will have real targets for the metrics that started as TBD.

The blocker is the story. File DLP blocking SharePoint search is the most important metric in the table — not because of the number, but because it explains why the flagship capability isn’t live yet. A sustainability-informed PM reports the blocker, explains why it exists, and shows what’s being done. That’s more useful than a status report that leads with the wins and buries the constraint.

Training completion is a leading indicator. 98.1% training completion before expanding access is a governance win. It’s also a sustainability win — tools adopted without training create technical debt, support burden, and misuse risk. The training-first approach costs time upfront and saves significantly more later.

What’s Next

The next P5 update will report:

  • File DLP completion (AvePoint) and the SharePoint search milestone
  • Updated meeting hours saved target based on trajectory
  • Prosperity target based on the 28K actions/month trend
  • Power Platform support model (due Q3)

I’ll also be adding metrics as the program matures. More data points are coming — particularly around AI governance and the Power Platform environment strategy.

The Broader Point

If you’re a PM who has dismissed sustainability metrics as soft, aspirational, or irrelevant to IT projects — this is what they look like when applied to a real program with real data.

People, Planet, Prosperity, Process, Products. Five dimensions. Measurable from the start. Reported honestly, including the blockers.

That’s not a compliance checkbox. That’s how you know if a program is sustainable — not just functional.

This is part of my 2026 public goals series. I’m documenting the journey from planning through results — including what doesn’t go as expected.

Related posts: