Deep Dive Team 26 - AMP
Atlassian AMP: AI Is Joining the Team, But Is Your Business Ready?
At Team ’26 Europe, Atlassian introduced AMP, its vision for humans and AI agents working together. But what does this actually mean for businesses, and how can organisations prepare?
If you’ve been following Atlassian’s announcements recently, you’ll have noticed one very obvious theme: AI is becoming an increasingly important part of how we work.
We’ve moved from AI helping us write content and summarise information to Rovo agents supporting tasks, searching across systems and assisting with workflows. At Team ’26 Europe in Amsterdam, Atlassian took things a step further with the announcement of AMP, the Agentic Multiplayer Protocol.
Now, before you switch off at the mention of another acronym (because the technology industry clearly doesn’t have enough of those), this one is worth understanding.
AMP represents a shift in how Atlassian sees the future of work. Rather than AI being something we interact with individually, the vision is for AI agents to become active participants in our teams. It sounds promising, but what does that actually mean for businesses? And how do we embrace these capabilities without compromising security, accountability or the processes we’ve spent years developing?
What exactly is AMP?
In simple terms, AMP is Atlassian’s approach to enabling people and AI agents to collaborate within the same working environments, using shared information, defined permissions and visible workflows.
Think about how most of us use AI today. We open a chat, ask a question, generate some content and perhaps copy the result into Jira or Confluence. That interaction is largely between us and the AI. AMP aims to change this by bringing AI agents directly into the work we’re already doing, allowing them to contribute to tasks, discussions and workflows alongside human colleagues.
Atlassian describes this as moving from single-player AI to multiplayer AI. Importantly, AMP isn’t simply another chatbot or automation tool. It’s a broader framework designed to make collaboration between people and AI agents more connected, visible and manageable.
What could this mean for everyday business operations?
Let’s forget the technical terminology for a moment and look at how this might work in practice. Imagine a customer raising a support request through Jira Service Management. Today, an agent might need to review the request, search for similar incidents, consult documentation and decide what happens next.
With appropriately configured AI agents, parts of that process could be supported automatically. An agent might identify missing information, locate relevant knowledge articles, suggest a classification or prepare a response for human review.
Or consider a project manager coordinating several teams. Rather than manually reviewing Jira boards, chasing updates and checking Confluence documentation, an AI agent could help gather progress information, identify blockers and prepare a project summary.
The opportunity isn’t simply about completing tasks faster. It’s about reducing the time people spend searching for information, repeating administrative activities and switching between systems. And because AMP focuses on collaboration within existing workflows, AI contributions can become more visible to the wider team.
Of course, these are examples of what agent-assisted working could look like, rather than capabilities every business can immediately switch on.
The business benefits, and the reality check
There are some clear opportunities for organisations already using Atlassian products. AI agents could help reduce repetitive work, improve access to organisational knowledge, support more consistent processes and free employees to focus on activities that require human judgement. But there’s something businesses need to consider before getting carried away.
AI won’t automatically fix inefficient processes.
If your Jira workflows have accumulated years of unnecessary complexity, your Confluence documentation is outdated or your permissions are poorly managed, introducing AI won’t magically resolve those problems. In some cases, it could make them worse.
An agent relying on conflicting documentation could provide incorrect guidance. One operating within an overly permissive environment could potentially access information it doesn’t need. AI can be incredibly useful, but it still depends on the quality of the information, processes and controls surrounding it.
Or to put it another way, AI could simply help us make the same mistakes at a much more impressive speed.
Do businesses need to rethink their processes?
Potentially, yes. Most business processes have been designed around people performing tasks, making decisions and handing work to colleagues. Introducing AI agents means reconsidering where responsibilities sit.
Take change management as an example. An AI agent could help prepare change requests, check required information and identify similar historical changes. But should it approve a change? Execute it? Decide whether the associated risk is acceptable?
Those decisions need careful consideration. For many organisations, the sensible starting point will be allowing AI to assist with gathering information and recommending actions, while retaining human approval for activities carrying significant operational, financial or security risks.
The objective shouldn’t be to remove people from every process. It should be to make better use of their time and expertise.
Security and governance: the conversation we can’t ignore
One of the more encouraging aspects of AMP is Atlassian’s focus on agent identity, permissions and visibility. The intention is that agents can have identifiable profiles, defined ownership and traceable contributions.
This matters. If an AI agent updates a Jira issue, changes documentation or contributes to an important business decision, organisations need to understand what happened and who is accountable. But having governance capabilities available doesn’t automatically mean an organisation is governed effectively.
Businesses should be asking some important questions:
What information can our AI agents access, and do they genuinely need it?
Which activities can agents perform independently, and which require approval?
How will we monitor agent activity and investigate mistakes?
Who is responsible for managing agents throughout their lifecycle?
Are our existing security policies and data protection controls suitable for AI-assisted working?
These questions become particularly important when AI interacts with customer information, employee records or business-critical systems. Atlassian is developing additional capabilities around agent identities, non-human identity management and AI governance, but organisations will still need to assess which controls are available and appropriate for their environments.
An AI agent might be capable of performing a task, but that doesn’t necessarily mean it should have permission to do so.
Preparing your organisation for AMP
My recommendation would be to treat AMP as an opportunity to review how your organisation works, rather than another feature to enable as quickly as possible.
Start by reviewing your existing Atlassian environment. Are your workflows still fit for purpose? Is your documentation accurate? Are permissions appropriately managed?
Then identify where AI could genuinely add value. Look for repetitive, time-consuming activities that could benefit from assistance, and begin with a controlled pilot rather than introducing agents across the business immediately.
Establish clear ownership and governance from the beginning. Decide what agents can access, which actions require approval and how their performance will be monitored.
And don’t forget the people involved. Employees need to understand how to work with AI, recognise incorrect outputs and know when to question or escalate a recommendation.
Finally, measure the results. Saving time is useful, but not if it comes at the expense of accuracy, customer experience or additional rework.
My thoughts: AI readiness matters more than AI adoption
Having followed Atlassian’s developments across Jira, Confluence, Rovo and the wider platform, I think AMP represents an interesting evolution in its approach to AI.
What stands out to me isn’t simply the introduction of more AI functionality. It’s the emphasis on bringing agents into existing teams and workflows, with greater visibility and accountability. That’s an important distinction.
However, I don’t think organisations should see this announcement as a reason to rush into widespread automation. Some of the capabilities supporting AMP are already available, while others are still being introduced or remain in early access.
Businesses need to understand what is available to them, where it can provide value and whether their existing processes are ready to support it.
Ultimately, I believe the organisations that benefit most won’t necessarily be those introducing the largest number of AI agents. They’ll be the ones taking the time to understand their processes, establish appropriate controls and identify where AI can make a meaningful difference.
The future of work may well involve humans and AI agents working side by side. The challenge is making sure they’re working together effectively, securely and with a clear purpose.
And before we give AI a seat at the table, perhaps we should make sure it knows which meetings it actually needs to attend.
How Sourcesense can help
At Sourcesense, we help organisations get more value from their Atlassian environments, from reviewing existing configurations and improving workflows to supporting the adoption of new capabilities.
As AI becomes more integrated into the Atlassian platform, having the right foundations in place will be increasingly important. Whether you’re exploring Rovo, reviewing your Jira and Confluence environments or considering AI-assisted workflows, our team can help you understand your options and prepare for what’s next.
Want to explore what AMP and Atlassian’s latest AI developments could mean for your business?
Comments