Cloudwell Conversations, Part 1: ShareGate’s Richard Harbridge on Migration in the AI Era and Why You May Need to Move the Mess Before You Fix It

This summer, SharePoint Server 2016 and 2019 finally ran out of road. End of support, July 2026. If your migration plan was “get everything into the cloud, tick the box, never speak of it again,” Richard Harbridge has some news for you.

The finish line has moved.

ShareGate’s 2026 survey of 943 IT professionals found migration is nearly universal: 94% moved something in the past two years, most juggling two or three migration types at once, and for nearly a quarter the biggest job was a merger, acquisition, or divestiture. And AI has turned what used to be a one-way trip into a cycle: archive, consolidate, reorganize, repeat.

Canada-based Harbridge has watched this from every angle. He built and led consulting businesses on two sides of the border, spent two years at Microsoft working with partner executives in the early Nadella era, and is now Principal Industry Advisor, Microsoft 365 at ShareGate, where he does everything from leading CIO think tanks to helping interpret and communicate ShareGate’s Microsoft 365 research into how organizations actually manage Microsoft 365.

His argument is simple: migration isn’t a project you finish anymore. It’s a capability you build.

He joined Cloudwell Conversations to explain what that means in practice and why the smartest thing you can do with a messy file share might just be to move it exactly as it is.

You’ve said migration is never really finished. What’s changed?

Richard Harbridge (RH): A lot of organizations still think of migration as legacy-to-leading-edge modernization, move as much as you can off the old platform, unlock new potential, done.

But in the AI era especially, there’s a lot of value in moving your content around again. Once more content is connected within the same cloud environment, AI can surface more of what each user already has permission to access. That makes the quality, ownership, permissions, and organization of the underlying content much more consequential. So, people archive a bunch of stuff, because no one deletes anything (if we’re being honest), or they move things into containers that AI prioritizes differently.

There’s this ongoing optimization and reorganization work. I’d classify a lot of that as migration. We just don’t call it migration. I frame it as consolidation.

Sprawl occurs, and sprawl compounds like interest. At some point there’s so much of it that even good governance isn’t really managing it, and you have to consolidate, merging sites and teams, archiving, moving active content around. Add M&As and divestitures on top, and tenant-to-tenant consolidation has become one of the most consequential migration patterns we are seeing.

Any organization that asks “how do we repeat this, so the next time we migrate we do it better?” is going to be in a much better position. Especially because I expect continued M&A, divestiture, and consolidation activity as organizations restructure around AI and changing market conditions.

“Sprawl compounds like interest. At some point, even good governance isn’t managing it.”

Why are organizations migrating right now?

RH: Security still leads, 45% saying this was the biggest motivator in our latest survey. End-of-life SharePoint Server is a great example: in the AI era, once support ends, those legacy systems can become a growing security, support, and operational risk.

Scalability is next at 42%, and that’s partly AI, they want access to modern AI capabilities, and for Microsoft 365 Copilot and many related experiences, that generally means moving content into cloud-connected services.

Compliance is 39%, much higher than last year, and I think that’s because compliance is also becoming more strategically relevant. Leaders increasingly recognize that it shapes how confidently AI can be deployed and scaled.

And AI readiness itself is now a migration driver in its own right, for over a quarter of organizations, it’s the reason they’re migrating now rather than later.

We saw it in the telemetry first. We expected on-premises-to-cloud migration volumes to continue declining as the remaining installed base became smaller. Instead, we saw a renewed increase. When we examined the reasons, AI readiness emerged as an important contributor. You’d think everyone who really needed AI had already moved. No, there are whole industries just entering that enablement phase now.

What does good pre-migration planning actually look like?

RH: Three things. First, understand what you have. Organizations that can say “I have 10,000 sites, and here are the departments, the owners, the metadata”, not just “I have 10,000 sites”, are in a far better position.

If you’re doing that inventory anyway, use it. A migration is a great catalyst for governance: you’ve identified the content owners for verification and UAT, so keep that information. Don’t throw it away when the project closes.

Second, identity. In our research, 36% of organizations said identity preparation was the single hardest part of a migration, harder than permissions, harder than the content itself. It’s the part that gets bolted on last.

On the surface a SharePoint or Teams migration looks like it’s about the content, but you’re often moving tenant to tenant or on-prem to cloud, and identity is the basis for everything else. Get the mapping wrong and you’re reworking the whole thing later.

Third, get help early and tool up early. 82% of organizations used a partner during M&A migrations, and the ones who brought the partner in at the planning stage were the most satisfied afterwards, because a good partner asks the questions that surface complexity you’re not thinking about. And 30% of people told us their top regret was not investing in purpose-built tooling sooner. Native capabilities should always be assessed and used where they fit, but they may not provide enough automation, reporting, or control for complex migrations at scale.

“The hardest part of a migration is usually the part that gets bolted on last. This year, that’s identity.”

Nobody wants to delete anything. Why not, and what should they do instead?

RH: Archiving is way up, and that’s genuinely exciting. Microsoft 365 Archive can be one of the most practical ways to improve Copilot relevance because archived SharePoint content is not used in Microsoft 365 Copilot responses, while Purview, retention, and eDiscovery capabilities continue to apply. Reactivating archived SharePoint content is now free, although newly reactivated content cannot be re-archived for four months. Organizations can archive entire SharePoint sites and, more recently, individual files, although file-level archiving still has some important limitations.

The reason people won’t delete is visibility. Many traditional file shares do not provide reliable last-viewed signals without additional auditing or analytics, so teams often lack the evidence they need to delete confidently. So, people understandably hang on to everything, because they don’t know. But once that data has sat in SharePoint for three to six months with zero views, no one’s even breathed on it, they get comfortable: archive it, see if anything breaks, see if anyone shouts. Then delete later. I’d suggest that around 95% of the time it’s an archive event, not a deletion event and the deletion happens on the archive afterwards.

“When visibility is low, archiving is often the bridge between keeping everything and deleting with confidence.”

When the content actually moves, is it big bang or phased?

RH: Everything is more connected than it used to be. It’s not just free/busy sharing anymore. I share a link to an asset and expect it to keep working; meetings have assets inside them. So even when you phase, each phase tends to be a more complete, more comprehensive migration than it would have been five years ago.

But you still phase, because there’s work in validating, verifying, prepping. And pilot. In my experience, the migration projects with the highest success and satisfaction consistently include a well-designed pilot. It’s the number one thing. Every time, the pilot surfaces the assumptions that were off.

“I’ve never seen a migration with high success and satisfaction that didn’t have a good pilot.”

What’s the biggest migration mistake you see?

RH: Conflating a migration project with an optimization project. They’re two different budgets, two different styles of project, managed differently.

The migration itself is usually a bounded project: the source, destination, volume, and acceptance criteria can be defined. It is often IT-led, although business validation, compliance decisions, and change management still matter. Optimization is different because it requires deeper business participation and has no natural endpoint.

Here’s the thing most people get backwards. Say you’re migrating a traditional mess of folders from a file share. You’re tempted to restructure it before you move, get the house in order first. But your restructuring tools in a file share are terrible. You can’t even see view dates. In the cloud, you have AI tools. Move the mess as-is, then use something like autofill columns, a prompt per document, extracting lease terms or customer type or region as metadata, cheaply and accurately. Now the restructuring is easy: grab everything matching this metadata, move it there. You’ve done it faster, for less, and you’ve taught your people a metadata skill that agents will consume later too.

So about 90% of the time: lift the mess, optimize in the destination. The exceptions are foundational stuff, classic SharePoint sites to modern containers, obvious junk and test data. One of my favorite patterns is migrate-then-archive: migrate it, verify it landed, archive it in place using Microsoft 365 archiving. eDiscovery works, the technical team has succeeded, and if nobody touches it, your disposition process deletes it later.

“90% of the time: lift the mess, optimize in the destination.”

So, migration becomes something you’re good at, rather than something you survive?

RH: Exactly. Think of it as a flow you budget and plan for, not a one-off. The organizations asking “how do we repeat this better?” are the ones building the muscle and all that inventory, attribution, and optimization work is the same work that makes your AI and governance journey successful. Which is where the real story starts.

Richard Harbridge is Principal Industry Advisor, Microsoft 365 at ShareGate and a long-standing Microsoft MVP. A business analyst at heart, he built and sold consulting businesses in the US and Canada and spent two years at Microsoft working with partner executives before joining ShareGate, where he leads CIO think tanks and the company’s annual Microsoft 365 research. Find him on LinkedIn.

Planning a File Share, SharePoint or tenant-to-tenant migration? Cloudwell has been moving organizations to Microsoft 365 for decades and we start with the planning questions that save you pain later. [CTA: Contact us for a consultation]

Other useful blogs from Cloudwell on migration:

https://cloudwell.io/sharepoint-copilot-readiness-box-dropbox-migration/

https://cloudwell.io/sharepoint-2016-2019-migration-readiness-checklist/

https://cloudwell.io/sharepoint-2016-2019-to-sharepoint-online/

https://cloudwell.io/your-sharepoint-migration-isnt-a-technical-project/

https://cloudwell.io/sharepoint-2016-and-2019-end-of-life-cloud-migration-options-for-it-leaders/

https://cloudwell.io/sharepoint-2016-2019-end-of-life-on-premises/

In Part 2, Richard explains why the day you stop migrating is the day you start governing and why 93% of the people who run Microsoft 365 think their governance is in good shape. The incident data tells a more complicated story.