In 2009, Diego Domingos da Silva sat in a job interview for a SharePoint developer role and told the panel he had never used SharePoint.
Their answer? That’s OK, neither have we. Figuring it out would be his job.
He hasn’t left the Microsoft collaboration stack since.
Seventeen years on, he is a Microsoft 365 Solutions Architect, a 2x Microsoft MVP for SharePoint and Microsoft 365 Copilot, and a Microsoft Global Community Initiative Regional Leader. He also writes Unsuck M365, a blog with a simple premise: the technology isn’t bad once you take the time to understand it.
He has watched organizations fight the cloud for years, then move onto Teams almost overnight when the pandemic left them no choice. AI hasn’t given anyone that much time. Employees are already using it, IT is still writing policy, and Copilot is surfacing every overshared site and broken permission that teams have spent years stepping around.
Every one of those problems was there before Copilot arrived. What’s changed is how visible they are.
He sat down with Cloudwell to talk about what really stalls a Copilot rollout, what to do when adoption flatlines, and where in-house IT teams should start when everything is happening at once.
You landed your first SharePoint job without knowing SharePoint. How did that turn into Unsuck M365?
That was 2009. I said during the interview, “I don’t have experience with SharePoint.” And they said, “It’s OK, because neither do we. It’s going to be your job to come here and figure out the platform.”
Back then people were still using network drives. So why would they put their files into this weird website we were telling them was the way to go?
There was a lot of “SharePoint sucks.” That’s when I started saying, well, let’s unsuck this together.
I didn’t want to tie it to SharePoint, because the technology changes so much. So, I called it Unsuck M365. It talks about the things people complain about, but really it’s about doing it better together. If you understand the technology, the technology is not bad. You just have to spend some time understanding what it is.
What were people complaining about back then, and what changed?
Mostly user experience. People said it was ugly. I lost count of how many times someone asked me, “Can you make it not look like SharePoint?”. Often that got in the way before people even spent the time to learn the technology.
Then Microsoft made a huge investment in the look and feel. Now the sites look amazing. These are sites people are proud of.
But it wasn’t only the looks. The pandemic pushed everyone onto Microsoft Teams, and that brought people even closer to SharePoint without them even knowing. Tighter integrations helped too. So we went from people being so anti-SharePoint to it being something they can’t live without.
You’ve watched a lot of resistance come and go. What has it taught you about adoption?
Cloud was probably the biggest resistance I’ve seen. For decades, organizations thought control meant owning a server in a room. The idea that their information would no longer live on their own servers was a hard sell for years.
Then Teams came, and it was probably the fastest adoption we’ve seen. Organizations that could spend years debating a SharePoint migration were almost forced onto Teams. Everyone was working from home. What else were they going to use?
Both taught us the same lesson. People rarely adopt technology because IT tells them to. They adopt it when it solves a problem they really care about.
So it’s about bringing the technology to where people work, not shoving it in their faces.
Not everyone is as excited about this stuff as we are in the MVP community. Most people just want to get their work done, and get it done fast.
Where does AI fit into that story?
Go back a couple of years and we were having arguments with users about metadata, about tagging their files. Governance was hard because you had to explain to people: do this, don’t do that. People would do it for a while, then work got in the way, they got busy and they stopped.
Now you can automate a lot of that. AI fills in the gaps for the things users find difficult to do on their own.
Copilot takes it even further, because it comes to you through meetings, emails and documents. That changes the architecture conversation.
Back then, a badly organized SharePoint site was annoying. Now you get poor results when you ask Copilot something, because your site isn’t organized. So, people have a lot more reason to get their content ready.
What does the pace of change look like for in-house IT teams right now?
One word: chaotic.
Cloud took years. AI arrived in months. With cloud, organizations had all that time to argue about it, see what worked and what didn’t.
Well, guess what? AI isn’t waiting for any of that.
Employees are using it whether IT likes it or not. Meanwhile, IT is still out there trying to write policy.
So, governance, security, information architecture and adoption are all happening at the same time.
And AI has this ability to expose every piece of technical debt you’ve been ignoring. It’s like shining a massive spotlight on everything you’ve done wrong: overshared sites, bad permissions, duplicate content.
None of that is new. Copilot just makes it a lot harder to ignore and pretend it isn’t there.
So where do IT teams start?
By looking outside of where you’re standing. Roles have gotten a lot blurrier. Before, you had an admin, a developer and the information worker. Well, guess what? Information workers are now building apps.
So, everyone has to think about things outside their own area of responsibility, like governance and information security.
If you try to stay in the same lane and do the same work in the same way you’ve done it your whole career, you’re not going to be in good shape.
Empathy goes a long way. Understand how other people work and what their jobs involve. Just knowing your own job isn’t enough anymore.
What’s the mistake you see most often?
Businesses buying licenses before they’ve even decided what success looks like. I’ve seen that fail more than once.
People get Copilot, try a few prompts, get mediocre answers and decide that’s what it does. But they’re not doing is looking at where Copilot is pulling that information from, whether the sources are good, whether the permissions make sense, whether users have been shown the right scenarios, or whether leadership is using it.
The thing people often forget to talk about is organizational change management. Bring people along for the ride. If users are just told, “We’re using Copilot from November 1,” with no training, it won’t work.
Business training matters. Communication matters. It’s not a one-person game. Everyone has to collaborate.
Have user groups, power users, a tiger team, whatever your organization calls it. Have an internal committee reviewing the changes. And get agreement from leadership, because those groups are a complete waste of time if leadership isn’t included in the discussion.
What does a Copilot rollout that works look like?
The best approach I’ve seen over the years is to start small. Don’t treat a pilot as, “Here are the licenses, everybody, we do AI now, let’s go do AI.” Work with a group around the things they actually do.
As IT, we’ve been used to installing something, training users and walking away. That’s not how it works anymore. If you’re not bringing people along from day one, it’s not going to work. Understand their processes and their day-to-day work.
I’ve worked in organizations where it failed because the person buying the licenses just gave them out to leadership. Give them to the information worker. Understand what they do every day and how to cut down the time it takes them.
Not everyone cares about the technology like we do. Some people just want to go into the office, do their work and walk away.
Six months in, Copilot adoption has stalled and the finance director wants to see the value. What do you do?
Start with the people who stopped using it. Not the ones who never used it at all. The ones who started and then stopped.
Ask them: Why did you try it? What worked? What did you expect it to do? Why didn’t you come back? Work out whether you have an awareness problem, a training problem, a bad use case, an information problem, or licenses assigned to the wrong people.
Stalled adoption is a symptom. Figuring out what caused it is a much better tactic than throwing another training session at everybody.
If IT is about to hand out Copilot licenses, what should they do first?
Really understand what Copilot is going to be allowed to see. Before you give anyone a license, look at permissions, governance, oversharing, stale content. Copilot only respects your Microsoft 365 permissions. That doesn’t mean your permissions are already good.
I’ve said this many times: AI readiness, Copilot readiness, that’s just governance with lipstick. The problem is governance isn’t sexy. Governance doesn’t sell until something breaks. Call it AI readiness and suddenly people are listening.
If documents are coming out that people aren’t supposed to see, you don’t have a Copilot problem. You have a permissions problem. People can ignore permissions as much as they want but put AI on top and it becomes obvious.
AI won’t create good information architecture for you. It will just expose every information architecture problem you already have.
***
Diego Domingos da Silva is a Miami-based Microsoft 365 Solutions Architect, a 2x Microsoft MVP for SharePoint and Microsoft 365 Copilot, and a Microsoft Global Community Initiative Regional Leader. He has worked in the Microsoft stack since 2009, from WSS and MOSS through SharePoint Online, Teams, Power Platform and Copilot. He writes Unsuck M365. You can meet him later this month at the Power Platform Conference in Las Vegas and TechCon in Dallas, and at M365 Atlanta in November, or connect with him on LinkedIn.
Planning a Copilot rollout, or trying to restart one that has stalled? Cloudwell’s Copilot Enablement Program starts by working out what Copilot can see, then helps your people find the use cases that stick.