15:20Machine-translated from Azerbaijani by AIRead in Azerbaijani
Atlassian Unveils AMP Protocol Connecting People and AI Agents
Atlassian wants to bring company data together in a unified system for employees and AI agents that perform tasks independently. CEO Mike Cannon-Brookes announced the plan at the Team Europe 26 event in Amsterdam. According to him, for a powerful AI model to deliver useful results, it must understand the company’s previous decisions and work practices. Atlassian plans to turn its platform into the main collaboration hub providing access to this information.
A key part of the new approach is giving AI the overall content of the work—that is, its context. Cannon-Brookes said that if AI does not know why a previous decision was made, it cannot provide the necessary assistance with subsequent tasks. It is impossible to take full advantage of AI that remains isolated in a separate system. By connecting different types of information, the platform will help employees find the answers they are looking for more quickly. According to the CEO, agents do not replace people; instead, they should be viewed as new members joining the team. Under this approach, AI is not merely an individual tool; people and agents work together. The company believes that as employees and agents work faster, numerous decision-making and learning processes occur simultaneously. If these are not properly coordinated, the fast pace of work can cause confusion within teams.
Atlassian also unveiled a new collaboration protocol called AMP at the event. The system is designed to enable people and AI agents on various platforms to work together, both within Atlassian and in external services. AMP works with the new Atlassian MCP server. The server gives agents access to the company’s tools and data so they can complete tasks. One of the innovations demonstrated was the Loom Record for Agent feature. Instead of preparing a written request in Loom, a user can record a video and explain the details of the task. The agent then treats the video as a request. Atlassian wants to accelerate collaboration between people and multiple AI agents on complex tasks. However, the people making decisions will not be left out of the process.
A key part of the new approach is giving AI the overall content of the work—that is, its context. Cannon-Brookes said that if AI does not know why a previous decision was made, it cannot provide the necessary assistance with subsequent tasks. It is impossible to take full advantage of AI that remains isolated in a separate system. By connecting different types of information, the platform will help employees find the answers they are looking for more quickly. According to the CEO, agents do not replace people; instead, they should be viewed as new members joining the team. Under this approach, AI is not merely an individual tool; people and agents work together. The company believes that as employees and agents work faster, numerous decision-making and learning processes occur simultaneously. If these are not properly coordinated, the fast pace of work can cause confusion within teams. Atlassian also unveiled a new collaboration protocol called AMP at the event. The system is designed to enable people and AI agents on various platforms to work together, both within Atlassian and in external services. AMP works with the new Atlassian MCP server. The server gives agents access to the company’s tools and data so they can complete tasks. One of the innovations demonstrated was the Loom Record for Agent feature. Instead of preparing a written request in Loom, a user can record a video and explain the details of the task. The agent then treats the video as a request. Atlassian wants to accelerate collaboration between people and multiple AI agents on complex tasks. However, the people making decisions will not be left out of the process.