AI interaction guidelines
AI interaction guidelines are rules and best practices for how to design AI interactions.
Make workflows proactive
AI experiences should demonstrate an understanding of the context of the person, team, and company knowledge. Using that context, AI should make smart assumptions about when to be proactive and how to help achieve high-quality outcomes.
Keep the flow state
AI should preserve a user’s flow by keeping interactions natural and unobtrusive, with controls embedded directly in the working surface to reduce context switching. When transitions occur across interfaces or devices, tasks and state should carry over seamlessly.
Content and form are dynamic
Whether it’s text, audio, video, or visuals, AI should let people input or transform information in the form that best suits them. Interfaces should adapt just as fluidly, shifting controls and views based on intent.
Make it easy to ‘pop the hood’
Give people clear insight into system state and how outcomes are produced, in real time and retrospectively. This deeper view should be easy to access across surfaces, with options like logs, documentation, advanced controls, or technical layers, adjusted to user role and preference.
Promote collective intelligence
Connect individual work to the broader team’s knowledge, context, and workflows. Interactions should seamlessly surface relevant expertise, resources, and shared information so people can build on each other’s work while maintaining privacy and permissions. In doing so, AI amplifies both personal productivity and team impact.
Expand to multi-player and multi-intelligence
Support multiplayer collaborations by letting work move fluidly from individual effort to shared human-AI teamwork. This should feel continuous and connected across people, teams, and intelligent agents.
Weave a bit of Atlassian into the experience
AI should weave a subtle blue thread through every interaction, reflecting Atlassian’s brand through tone, visuals, behavior, and system response without overwhelming the experience. These cues, whether colors, micro-animations, word choice, or design assets, create continuity and craft while allowing stylistic and functional variation. Even in personalized experiences, these touches should anchor the interaction in a cohesive Atlassian identity.
Always accelerate responsibly
AI outputs should be high quality and efficient, not rushed or careless. Even when tasks are delegated, the system should clearly communicate risks, options, and opportunities to course-correct along the way.