I focus on sharing and finding better ways to work with technology: Microsoft MVP, strategist, speaker, author, and super-friendly advisor to hundreds of orgs.
If you remember nothing else about AI strategy, remember this:
Clarity first.
Confidence next.
Connection always.
Clarity about what you are doing.
Confidence from small real wins.
Connection to the people whose work is changing.
AI can make feedback easier to draft, but it cannot have the difficult conversation for you.
Leadership still requires sitting with discomfort, listening fully, and taking responsibility.
The more AI content you create, the more important content lifecycle becomes.
Reviews.
Expiry dates.
Owners.
If everything lives forever, people will not know what to trust.
It feels like Microsoft 365 governance should be getting easier.
Better first-party tools.
Better ecosystem tooling.
Better admin skills.
Better patterns.
But the surface area keeps expanding:
Content.
Guests.
Apps.
Automations.
Agents.
Digital worker identities.
AI workflows.
The water level has risen.
So has the size of the ocean.
AI governance is not shrinking.
It needs to become more repeatable, automated, and shared.
Many organizations already have the data they need for valuable AI scenarios.
What they lack is alignment on which problems to solve first and who will own the change.
Some of the biggest barriers to AI are actually about identity.
โI was always the person people came to for this.โ
When a tool can now help with that thing, it can feel like a loss.
Name that. Support that.
A comforting AI governance story:
We clean up the debt.
Then the workload drops back to normal.
I do not think that is how this works.
Even after cleanup, governance stays heavier because it is not just reducing AI risk.
It is driving AI success.
Better content.
Better ownership.
Better lifecycle.
Better trust.
Better outcomes.
Governance is not just the cleanup crew.
It is part of the AI performance engine.
Not all AI risks are equal.
Some relate to individual privacy.
Some relate to customer trust.
Some relate to intellectual property.
Categorising them helps you respond appropriately.
A lot of Microsoft 365 oversharing starts as a reasonable collaboration need.
โCan we share this with more internal people?โ
Then someone uses โEveryone except external usersโ at the site level.
The intent was narrow.
Permission inheritance made it broad.
That is the governance mismatch.
In the AI era, buried oversharing becomes much easier to discover.
Microsoft 365 sprawl is not just too many Teams or SharePoint sites.
It can be permissions, sharing, configuration, apps, automations, agents, change, licensing, content, devices, and workspaces.
AI readiness touches all of it.
https://t.co/WoKhU4NDDt
Sometimes the most strategic AI decision is about scope.
You may choose to focus on one business unit, one geography, or one set of processes first.
Depth beats thin coverage across everything.
Some of the best AI use cases start as โpersonal hacksโ that people build for themselves.
The opportunity is to spot the good ones and scale them carefully across the team.
Governance is a team sport.
Without roles and playbooks, everyone chases the ball.
With the right structure, talented people can move faster together.
That is how I think about Microsoft 365 governance in the AI era.
https://t.co/WoKhU4NDDt
A useful AI audit question:
โWhere are we using AI today that is not officially documented anywhere?โ
The answers can surface shadow AI in teams and tools you did not expect.
The credit card analogy for AI governance is useful:
Sprawl is the balance.
AI changes the interest rate.
Agents may add new charges.
At some point, the minimum payment approach to Microsoft 365 governance stops working.
https://t.co/WoKhU4NDDt
Sometimes the bravest thing a leader can say in an AI meeting is โI do not know yet.โ
Leaders often underestimate how powerful it is to say, โI am learning this too.โ
That sentence can remove a lot of pressure from teams trying to figure out AI, Copilot, and agents in real time.
The key is when we share "I do not know yet." to follow it with โHere is how we will find out.โ
AI and automation can create new blind spots.
If you automate a step, it can quietly disappear from peopleโs mental model of the process.
Make sure someone still understands what the agent is doing and why.
Not everything should be migrated first.
Sometimes the smarter move is to connect first.
The real question is not just โhow much data?โ
It is โhow entangled is the ecosystem around it?โ
#Microsoft365#M365Migration#Microsoft365Copilot#Copilot
Governance used to be treated like the brakes.
In the AI era, I think it is becoming more like traction control.
Not there to stop the car.
There to help you move faster without losing control.
Discussed this on The Enterprise AI Show.
https://t.co/WoKhU4NDDt
Microsoft 365 already contains a huge amount of organizational memory.
AI just makes it easier to surface.
If that memory is outdated, duplicated, or wrong, Copilot will faithfully repeat it.
Cleaning up knowledge is not optional anymore.