Your Copilot case study that leads on features and hours saved no longer cuts it with the person who approves the spend. The business holds the budget now, and it wants proof of order-of-magnitude change, not a 10% efficiency. For MSPs selling Copilot licences and adoption services, that means rebuilding your case studies as business transformation stories, with AI as the method rather than the hero. Read on to find out what belongs in case studies now, from the Microsoft market-intelligence lead who spelled it out to a room of Microsoft Partners at the recent IAMCP event in Paddington, London.
A Copilot case study now needs four things: it should speak to the business buyer as well as the CIO, now referred to as the “fusion team”, it should lead with an order-of-magnitude outcome rather than time saved, and it should treat AI as the method rather than the hero. Plus, show the transformation story, including how you redesigned, measured, governed and drove adoption of the process.
The reason all four matter comes down to one shift. A case study whose headline number is “20% time saved on admin” has already lost the buyer who now holds the budget. That was the uncomfortable message delivered by Ferdi Van Huffel, who leads market intelligence for Microsoft across EMEA, delivered at a recent IAMCP event. His job is to read the analyst research and tell Microsoft Partners what the market is actually rewarding, and his verdict on the time-saving case study was not pretty. Here is what he said needs to change, and what to put in its place.
1. Speak to the business buyer, not just the CIO
The person who decides whether your case study lands increasingly does not sit in IT. Van Huffel put hard numbers on it: business buyers now control around three quarters of AI services spending, and 60% of those decision-makers sit outside IT. A case study written in the language of governance, licensing and configuration speaks fluently to a reader who no longer holds the purse.
“If you’re not talking to the business buyer, you are deciding to play in the 25%,” he said.
This does not mean writing the CIO out. Van Huffel was clear that deal quality rises sharply when the CIO is accountable, and rises again when IT is on the delivery team.
The case study you need speaks to the business outcome the leadership team cares about, and the controllable risk the CIO has to sign off. Lead with the outcome, and give the technical reader their reassurance lower down, rather than making it the whole story.
2. Lead with order-of-magnitude outcome, not time saved
The single biggest change to your case study is what you count as the result. A marginal time saving does not justify the disruption an AI rollout demands, and it will not survive the finance director who has to fund it. Van Huffel was scathing about the pilot that proves someone’s afternoon got a little easier.
“10% time savings just don’t cut it,” he said. What the market rewards instead is change of a different scale: “an order of magnitude change. 10x. This is really transforming my business.”
The way to show that is to measure the process, not the task. His illustration is the template your case studies should follow: “What used to take five people is now taking just one. What used to take five days is now taking one. What used to cost you X is now saving you Y, in real money, not made-up human labour costs.”
A business leader discounts notional “hours saved” because it rarely reaches the P&L. A cost line that actually fell, a process that genuinely needs fewer hands, a turnaround measured in days rather than weeks, those are results a CFO can defend. Build your case study around one of those, with a real before and after, and you have proof that travels.
3. Treat AI as the method, not the hero
Here is the shift most Partners find counterintuitive, especially when the whole point is to sell Copilot. The case study should not be about AI. It should be about what the customer’s business can now do, with AI as the mechanism that made it possible.
Van Huffel says: “AI is how we did it, not the headline.” The result is the headline. The technology is the footnote that explains it.
He drew the contrast with what he called the spectator, the Partner whose case study amounts to “I configured your Copilot licences for you. Boring,” he said. That is a story about a task completed, not a business changed.
The version that wins tells the reader what a team can now do that it could not before, and only then explains that Copilot was how. Same project, different hero. The customer’s outcome takes the top line; the AI does its work underneath it.
4. Show the transformation story: redesign, measure, govern, adopt
A strong Copilot case study is a story with an arc, not a single result frozen in time. Van Huffel urges Partners to show the journey and, crucially, how the value compounds as it goes.
His worked example moved through the tools in sequence: a team taught to use Copilot Chat to think and generate ideas, then Cowork to redesign the process itself, then Copilot’s agent-building to automate what was left. The point of showing the arc is that “value compounds” as you move along it, and once the first process is redesigned, the next one is easier and the return grows.
What makes the story credible, and repeatable, is documenting your method rather than just the outcome. Van Huffel was specific about what that method is: “redesigning the process, a specific process, measuring it, naming it, putting governance alongside it, and showing the adoption, and then compounding.” Read that as a checklist for your case study.
Name the process you rebuilt. Show the baseline you measured it against. Explain the governance you put around it, which is where the CIO gets their reassurance. Prove that people actually adopted it, not just that licences were switched on. Then show how you moved on to the next process and the value stacked.
A case study built that way sells something a feature list never can. It shows a buyer not only that you got a result, but exactly how you would get theirs. That is what turns a proof point into a pipeline.
So look at your last Copilot case study and ask whose language it is written in, what its biggest number actually proves, and whether AI is the hero or the method. If it reads as “we switched on some licences and saved a bit of time,” it is describing the spectator Van Huffel warned us about. The story you need, the redesigned process and the order-of-magnitude number, is almost certainly sitting in your delivery team’s heads. It takes the right questions to get it out, which is the part we can help you with.
FAQs
Our results are more like 15-20% improvements, not 10x. Are we shut out of this?
Not if you measure the right thing. Order of magnitude rarely shows up when you time a single task; it shows up when you redesign the whole process around it. A task that gets 20% faster is unremarkable. A process that used to need five people and now needs one is a different order of result, even if each individual step only improved a little. Measure the redesigned workflow end to end, and show the compounding across several of them, rather than clocking one task in isolation.
Our client won’t let us publish hard financial figures. Can we still write an outcome-led case study?
Yes, because money is not the only order-of-magnitude proof. Capacity and process metrics carry the same weight and are often easier to clear: five people down to one, five days down to one, a backlog cleared, a turnaround that no longer needs overtime. Van Huffel’s own examples lead with those before they reach the money. Agree what you are allowed to measure with the client before the project starts, and you will have a defensible number even when the pound sign stays private.
Isn’t AI the whole point of a Copilot case study? Why downplay it?
Because the buyer isn’t purchasing AI, they’re purchasing the result it produces, and leading with the technology marks you as the Partner who just sells licences. “AI is how we did it, not the headline” is not modesty, it’s positioning. The outcome is what a business leader acts on and what their AI can cite when it summarises you. Copilot still features prominently, as the method that made the outcome possible. It just doesn’t get top billing over the result.
***
Need help writing your Copilot case studies? We’ll interview your experts to find the outcome, write the story, and build it into a campaign you can put
Co-op-funded paid spend behind. Get in touch and let us help you find your first one.
