Why Microsoft Copilot projects add absolutely no value

…unless you do the boring bits.

Ah yes, Microsoft Copilot. The magical AI assistant that will solve all your business problems overnight. Finally, a chance to automate mediocrity at scale! Management loves it, IT doesn’t want to hear about it, and users think it will do their job for them. Everyone has swallowed the Microsoft red pill! Spoiler alert: it won’t.

There are three main reasons to why these projects fail, so let’s unpack why most Copilot projects are doomed to add precisely zero value for the customer.

1. Leadership (Or lack thereof)

Apparently, when something shiny and new comes along, management believes the best course of action is to hand it straight over to IT. After all, if it plugs into Teams, it must be an IT thing, right? Wrong.

Copilot is not a project you can just throw into the basement with the servers and hope for the best. It’s about how people work, collaborate, and make decisions. You know – the messy human stuff. But instead of leadership stepping up, taking ownership, and driving adoption, they often retreat into the shadows, muttering something about “digital transformation” and waiting for the quarterly report to magically improve.

But hey, who needs leadership when you’ve got buzzwords?

How to fix it:

Leaders need to own Copilot adoption, not outsource it. That means clear goals, visible sponsorship, and a willingness to change how the organisation works – not just dropping Copilot in and hoping for miracles. Treat it like a business transformation, not an IT installation.

2. Training? What training?

But instead of acknowledging that training is a continuous, never-ending commitment, companies tend to roll out a single “lunch & learn” session, tick the box, and then wonder why adoption falls flat.

Here’s a hint: using Copilot effectively is not a one-off event. It’s a muscle you build. If you don’t support your people, you’ll get exactly what you paid for: nothing.

But sure, one PowerPoint session should be enough to completely change how people work… right?

How to fix it:

Make Copilot training a journey, not a checkbox. Create ongoing learning opportunities: role-based workshops, internal communities of practice, and real-life use case sharing. Celebrate the wins, share the flops, and keep at it. If you want adoption, you have to invest in your people, not just the licence fees.
If you have M365, you will have access to Viva Engage – USE IT!

3. Data, glorious data

And now to the drum I’ll keep banging forever: the value of data. If your data is a dusty attic of outdated documents, random duplicates, and “final_v4_really_final_this_time” files, then Copilot will happily serve that junk back to you with a bow on top.

Garbage in, garbage out. Copilot is only as good as the information you feed it. If you haven’t invested in cleaning, classifying, and actually moving your data into Microsoft 365, don’t expect AI magic. Expect frustration.

But please, tell me again how working from file shares and Outlook is “working just fine.”

How to fix it:

Sort your data. Full stop. This isn’t just about spring cleaning – it’s about making sure the right information is available, secure, and useful. Microsoft Purview is your best friend here: it helps you classify, protect, and govern data so that Copilot isn’t just serving up rubbish. Clean data means relevant answers, safe sharing, and insights you can actually trust. And make people delete all the stupid stuff… Be strict, yes it takes time, but it is soooo worth it!

Your data is your company’s value. Shame it’s buried in a digital landfill.

A bit about me (a.k.a. why I won’t shut up about this)

For the last two years, I’ve been working as a Technical Architect on Copilot projects inside different companies. And before that? Well, let’s just say I’ve been in this game long enough to remember when SharePoint was still considered cutting-edge, when file shares ruled the world, and when Teams was just that annoying new thing nobody wanted to use. (20+ years with SharePoint, 8–9 years with Teams… I’ve got the scars to prove it.)

And yes, I’ve always been that person droning on (and on, and on) about cleaning up your data, making it secure, and actually sorting it before you throw shiny tools on top. Guess what? That hasn’t changed with Copilot. If anything, it’s become even more critical.

So when I say your Copilot project won’t succeed without leadership, training, and clean data, I’m not just being dramatic. I’ve seen it. Repeatedly. And I’d rather save you the pain of learning the hard way.

Here is a whole set of posts about Purview and Copilot.. we have mentioned it before…

In Conclusion

So, are Copilot projects worthless? Absolutely not. But without leadership, without ongoing training, and without getting your data house in order, they’ll be about as useful as a chocolate teapot.

The technology works. The question is: will you?

Because let’s face it: Copilot can’t fix lazy strategy, bad habits, or your 17 versions of the same spreadsheet.