No, it does not.
Atlassian Rovo Chat now lets you share an entire chat thread with your colleagues.
This often raises the question:
Suppose I’m discussing a project with Rovo,
referencing the following in the conversation:
Jira tickets,
Confluence pages,
customer support data,
and internal project context.
Then I share the full chat with another colleague.
Does that mean they also get access to all the data I originally could see?
The answer is no.
Atlassian clearly states:
Shared chats remain permissions-aware.
In other words,
sharing a chat conversation
does not mean sharing permissions.
Share Chat Shares the "Work Context"
The new Share Chat feature in Rovo Chat
primarily addresses a common team issue.
Imagine you spent an hour discussing with Rovo:
Why the product launch is delayed.
During the conversation, you consulted:
Jira issues,
Confluence documents,
project decisions,
and past records.
Eventually, you come up with a plan.
Now, another colleague needs to take over.
In the past, there were usually two options:
Explain everything again,
or write a separate summary.
The problem with summaries is they often only contain the final answer.
The reasoning behind excluding certain options,
uncertainties remaining,
which directions were considered earlier,
are often lost.
Share Chat allows the colleague to pick up directly from this
conversation context
and continue seamlessly.
But Context Is Not the Same as Source Permission
It’s vital to distinguish two things here.
First:
What happened in the discussion.
Second:
Whether you have permission to view the original source material.
These are not the same.
For example, you might be a manager
with access to a private Confluence Page A,
while another colleague does not have permission.
You and Rovo’s chat referenced Page A.
When you share the chat with that colleague,
this action itself
does not transfer the permissions for Page A.
Atlassian’s official stance is very clear:
The information in Shared Chat still follows permissions,
and colleagues can only see
the information they were originally authorized to view.
So Share Chat Cannot Bypass Jira Permissions
Here’s a clearer example.
An engineering manager can see a
security incident Jira issue,
but marketing colleagues cannot.
When the manager discusses launch risk with Rovo,
this Jira issue could be part of the background data.
Then, the manager shares the chat with marketing.
This does not mean marketing can:
open that Jira issue,
view all its contents,
download attachments,
or browse other restricted issues.
Jira’s permission settings remain intact.
Share Chat is not a permission escalation tool.
The Same Principle Applies to Confluence
For example, a confidential acquisition plan
only accessible to the management team.
If Rovo uses this data in a conversation with a manager who has permission,
it does not mean sharing the chat will grant that page
to others without access.
Rovo’s core design always respects user permissions.
Atlassian Support also clarifies:
When Rovo performs searches and skills,
it only uses
content the current user has permission to access.
So who takes over the conversation
really matters.
The Same Question Can Yield Different Answers for Different Users
This is another aspect often overlooked in enterprise AI.
Suppose A and B both use Rovo.
A is a manager.
B is a regular employee.
Both ask:
“Why is this project delayed?”
If A has access to
financial documents,
management notes,
private Jira issues,
and B does not,
then the context Rovo uses differs.
Therefore,
the same prompt might be answered based on different data sets.
This is not a bug,
but rather the permission system working as intended.
What’s the Real Value of Sharing Chat Then?
If you can’t share all the original sources,
what’s the point of Share Chat?
The value lies in
avoiding restarting the entire thought process.
Colleagues can still see, within their permission scope,
the project background,
confirmed decisions,
action items,
discussion directions,
next steps,
and questions that remain unanswered.
So Share Chat effectively solves
context handoff,
not
permission handoff.
Remember these two phrases plainly.
What If a Colleague Can’t Access One of the Sources?
The correct approach is not
trying to bypass permission restrictions through chat.
Instead, ask:
Does this person really need that data to complete their task?
If not,
maintain the existing permission.
If yes,
go back to the source systems like
Jira, Confluence,
or the original data source
and formally adjust access rights.
Don’t assume that because
“I already shared the AI chat,”
permissions to the data should automatically follow.
The Same Applies to Agents
Rovo doesn’t just allow Share Chat.
It now also lets you directly @mention Agents within the chat.
For instance, mid-discussion,
bringing in a
Launch Planning Agent
or
Data Analyst Agent.
Agents can continue the current thread’s context,
but this does not grant the agent extra data permissions.
Atlassian’s principle for Rovo Agents is:
Agents themselves do not grant users additional data access.
Even if the agent is shared with another person,
that person can only access content they’re authorized to view.
So,
adding an agent doesn’t mean added permissions.
Why Is This Especially Important for Enterprise AI?
A common chatbot problem is simply:
Is the answer correct?
Enterprise AI has an even more complex challenge:
Whose data was used to generate the answer?
Because within companies, there are many sensitive information layers like:
HR data,
financial data,
customer data,
management documents,
legal files,
security incidents,
general project docs.
If AI automatically assumes
“Everyone can ask about any company data,”
then companies would never trust it.
Therefore, a critical enterprise AI capability is not:
Knowing all the company’s data,
but rather:
Knowing a lot of data while respecting what you are not allowed to see.
Correct Permissions Don’t Eliminate the Need for Checks
There’s another layer here.
Permissions-aware means:
Whether a user can access the data.
It does not guarantee that
Rovo’s summary is always accurate,
or that all shared decisions are current.
For example, yesterday’s chat might say:
“Launch is on Friday,”
but today the project plan changed to:
“Next Monday.”
Even with perfect permissions,
an old chat might still contain
outdated context.
So before sharing chat, it’s best to quickly confirm:
Have any decisions changed?
Are there any assumptions yet to be verified?
Which conclusions are just AI summaries and not official decisions?
This is a different risk than permissions.
The Easiest Way to Understand
Think of Rovo Chat as a work meeting.
You can hand over
the entire meeting transcript
to a colleague.
But this does not mean
every file referenced in the meeting
comes with access rights.
So when you see:
Share Chat
don’t automatically interpret it as
Share Everything.
The real rule is:
Chats can be handed off,
but permissions will not be upgraded.
Remember this one sentence:
Sharing context does not equal sharing access.
Today, grow a little with AI.
Learn one AI tip each day.
Save a bit of time every day.
Improve your skills daily.
SasaDaily, growing alongside you.
