Atlassian has introduced AMP, the Agentic Multiplayer Protocol, a new framework for coordinating work between people and AI agents across the Atlassian ecosystem. The company is positioning AMP as a shift away from “single-player AI,” where one person chats with one assistant, toward shared environments where agents can participate in team workflows with visible identity, context and accountability.
Atlassian announced AMP during Team ’26 Europe in Amsterdam. The company describes the protocol as a collection of experiences, technologies and operating patterns designed to help humans and agents collaborate inside tools such as Jira, Confluence and Loom, while also connecting to external systems through mechanisms including MCP and agent-to-agent standards.
What AMP is designed to solve
Most workplace AI adoption still happens in isolated sessions: one employee prompts an assistant, receives an answer and then manually moves that output into the rest of the team’s workflow. Atlassian argues that this model improves individual productivity but does not necessarily improve coordination across a team.
AMP is meant to address that coordination problem by giving agents a more explicit place inside shared work. Atlassian says the framework is built around three linked ideas: real-time and asynchronous collaboration, clear agent identity and presence, and context plus governance.
In practice, that means an agent may be briefed from a Loom recording, mentioned in a Confluence page, work against Jira issues, and continue a task across connected systems. Atlassian says actions should remain attributable so teams can see which agent acted, where it acted and what changed.
Why this matters for enterprise SaaS
The announcement is significant because it reflects a broader change in enterprise software architecture. AI assistants are moving from side panels into the workflow itself. Once agents can create, update and coordinate work across systems, the key product question becomes less about whether a SaaS platform has AI and more about whether it can safely orchestrate agents across shared business processes.
That creates a new layer of competition around identity, permissions, context and interoperability. SaaS vendors that own the system of record or the workflow graph may have an advantage because agents need structured access to people, projects, documents and decisions in order to act reliably.
It also increases the importance of governance. An agent that can move work forward asynchronously needs clear boundaries around who authorized it, what data it can access and how teams review its actions. Those are the same operational questions that show up when organizations move AI from pilots into production. XtraSaaS has covered that transition in our guide to moving AI pilots into production.
What Atlassian says AMP includes
Atlassian’s official AMP page describes three connected layers:
- Real-time and asynchronous collaboration: agents can work where teams already work and continue tasks across Atlassian and external tools.
- Identity and shared presence: agents have visible identity and attribution, allowing teams to see which agent is active and what it has done.
- Context and governance: agents can use shared context while organizations retain control over access, policy and oversight.
The company is also tying AMP to existing platform components including the Teamwork Graph, Atlassian MCP, its developer ecosystem and emerging agent-to-agent standards.
The SaaS strategy behind the protocol
For Atlassian, AMP is also a platform strategy. Jira, Confluence and Loom already sit close to the flow of work for many software and business teams. If agents become persistent participants in those environments, Atlassian can extend its role from collaboration software into coordination infrastructure for human-agent teams.
That is a meaningful distinction. A standalone AI assistant can answer questions, but an enterprise agent needs permissioned access to work objects, organizational context and action surfaces. Platforms that already connect those pieces are better positioned to provide that environment.
The same dynamic could influence the wider SaaS market. Products that expose reliable APIs, structured context and clear permission models may become easier for agents to work with, while closed products with weak interoperability may be harder to integrate into agentic workflows.
What SaaS leaders should watch next
- How open AMP becomes. The value of a protocol depends partly on whether third-party vendors and agents can participate without being tightly coupled to one vendor’s stack.
- How governance works in practice. Enterprise buyers will need auditability, policy controls and clear human accountability.
- Whether agents improve team outcomes. Faster individual tasks do not automatically translate into better cross-functional execution.
- How pricing evolves. Agent activity may create new consumption models around actions, orchestration or context access rather than traditional seats alone.
Sources and reporting
This report is based on Atlassian’s Agentic Multiplayer Protocol overview, the company’s Team ’26 Europe founder update, and Atlassian’s investor-relations news disclosures. XtraSaaS reviewed the primary-source material and added analysis focused on enterprise SaaS strategy, interoperability and governance.
