Microsoft Purview is bringing “last accessed” retention to SharePoint and OneDrive
One of the biggest challenges in Microsoft 365 isn’t necessarily creating information.
It’s getting rid of information we no longer need.
Over the years, organisations accumulate enormous amounts of content in SharePoint and OneDrive:
- Old project documents
- Historical reports
- Superseded policies
- Previous versions of presentations
- Outdated spreadsheets
- Former project team information
- Documents that haven’t been opened for years
And often, nobody knows whether any of it is still relevant.
Microsoft Purview Roadmap ID 472030 introduces an interesting new approach: the ability to apply retention to SharePoint and OneDrive files based on when they were last accessed.
The latest Microsoft Message Center update says Preview is expected to begin in late October 2026, with General Availability beginning in late December 2026 and completing by late January 2027.
And I think this could be much more significant than it initially appears.
🔒 We’ve traditionally thought about retention differently
Historically, retention tends to revolve around questions such as:
When was this document created?
When was it modified?
When was it labelled?
How long do we need to retain it?
Those are important questions.
But there’s another question that can tell us something very useful:
When did anyone last actually use it?
That changes the conversation.
A document created five years ago isn’t necessarily obsolete.
A document modified five years ago isn’t necessarily obsolete either.
But if nobody in the organisation has accessed it for five years, that’s potentially a very different story.
🚪 Introducing “last accessed”
With this capability, administrators will be able to use when items were last accessed as a retention control when configuring retention policies and labels for supported SharePoint and OneDrive content.
For example, an organisation might decide that certain content should be reviewed or removed when it hasn’t been accessed for a specified period.
This gives information governance teams another dimension to consider:
Not just how old is the information?
But:
How recently has the information actually been used?
💭 Imagine this scenario
You have a SharePoint site containing 250,000 documents.
Some were created last month.
Some were created ten years ago.
Some are actively used every day.
Others haven’t been opened since 2018.
Today, your retention strategy might struggle to distinguish between them if it’s primarily based on age or other metadata.
With a last-accessed approach, you can start to identify a very different category:
Information that has effectively become dormant.
That doesn’t automatically mean:
“Delete it.”
And this is an important distinction.
It means:
“This information hasn’t been accessed for X period — perhaps it should now be subject to a different retention decision.”
That’s much more interesting.
👍This could be a major information governance improvement
One of the biggest problems with retention is that organisations often retain information just in case.
Nobody wants to be responsible for deleting something that might eventually be needed.
So the safest decision becomes:
“Keep it.”
Then another year passes.
Then another.
Eventually, the organisation has millions of documents that nobody knows whether they still need.
This creates data hoarding.
And data hoarding creates problems.
More data means:
- More information to govern
- More information to classify
- More information to search
- More information to protect
- More information potentially accessible to users
- More information potentially available to AI
Which brings us to perhaps the most interesting part of this update.
🤝🏻 The Copilot connection
Microsoft explicitly says this capability can help improve the quality and relevance of Microsoft 365 Copilot responses by helping organisations remove obsolete data.
Think about that.
The problem isn’t simply:
“We have too many files.”
It’s:
“We have too much old information that may still be available when people – and AI – are looking for answers.”
If your organisation has years of obsolete documents sitting in SharePoint, they can potentially become part of the information landscape that users and AI interact with.
Good AI governance therefore starts with something much more fundamental:
Good information governance.
You can’t expect AI to produce consistently useful answers if your underlying information estate is full of outdated material.
🗝️ But be careful: old doesn’t mean useless
This is where governance teams need to exercise caution.
Completely obsolete.
Or it could be:
Extremely important.
Consider:
- A legal agreement
- A historical financial record
- A regulatory record
- A corporate governance document
- A long-term research record
- Evidence required for litigation
- A document subject to a legal hold
Nobody accessing the document doesn’t necessarily mean the organisation is allowed to delete it.
This is why last accessed should be treated as a governance signal, not an automatic deletion decision.
Retention requirements, legal holds, regulatory obligations and business requirements still matter.
ℹ️ The real opportunity: information lifecycle management
For me, this is where Roadmap 472030 becomes particularly interesting.
It potentially allows organisations to think about information as having a lifecycle:
Active
People are regularly accessing and using the information.
↓
Dormant
The information hasn’t been accessed for a significant period.
↓
Review
The organisation determines whether it still has business, legal or regulatory value.
↓
Retain or dispose
The information is either retained appropriately or disposed of according to policy.
That’s a much more intelligent information lifecycle.
🙈 It could also expose some uncomfortable truths
Imagine running this across your organisation and discovering:
40% of the documents in a particular SharePoint estate haven’t been accessed in five years.
That’s an interesting finding.
You might discover:
- Old project sites
- Abandoned collaboration spaces
- Duplicate content
- Historical information
- Orphaned documents
- Data that should have been deleted years ago
And that creates an opportunity.
Rather than simply buying more storage and accepting the ever-growing data estate, you can start asking:
Why are we keeping this?
That’s a much healthier information governance question.
💭 What should organisations start thinking about?
Even though Microsoft says no action is required before rollout, this is a good opportunity to review your current information lifecycle strategy.
I’d start by asking:
1. Do we know how much dormant content we have?
Not just how much data exists.
How much hasn’t been accessed?
2. Do we have retention policies that reflect business value?
Are we retaining information simply because we’re afraid to delete it?
3. Are our SharePoint sites being actively governed?
Do abandoned sites have an owner?
4. Could obsolete information be affecting Copilot?
Could old documents be making it harder for users – or AI – to find the information that actually matters?
5. What information must never be automatically disposed of?
Legal, regulatory and business-critical information needs appropriate safeguards.
💡 The Jim Talks takeaway
For me, the real significance of this update isn’t the new retention setting.
It’s the question it encourages us to ask:
“Is this information still useful?”
We’ve become very good at creating and storing information.
We’re becoming much better at protecting it.
But many organisations are still struggling with the third part:
Knowing when to let it go.
And as Microsoft 365 becomes increasingly AI-driven, this becomes even more important.
Because the future isn’t simply about having more information.
It’s about having the right information.
Information that is current.
Information that is trustworthy.
Information that is appropriately protected.
And information that is actually worth keeping.
Roadmap ID 472030 could be a small Purview feature with a surprisingly big information governance impact.
Leave a Reply