According to Atlassian research, most companies are approaching AI adoption backwards by optimizing how individuals use AI instead of how teams work together. Dr. Molly Sands, head of the Teamwork Lab at Atlassian, shared these insights during a fireside chat with VentureBeat at VB Transform 2026.
The gap between activity and value in AI usage
Atlassian's annual State of Teams Report, which surveyed 12,000 global knowledge workers and interviewed roughly 200 Fortune 1000 executives, found a significant disconnect between activity and value. 89% of executives reported that individuals are speeding up thanks to AI, yet only 6% could point to specific examples of clear ROI. However, about 14% of teams had translated AI usage into real value, meaning a single organization could contain high-performing teams surrounded by others seeing no return at all.
Three key factors: context, workflows, and culture
Leading teams share three characteristics. First, context: they build a context graph by capturing goals, decisions, and organizational knowledge in shared digital records rather than individual memory. Using products like Jira and Confluence, the graph connects work items, goals, and people, giving AI access to the organizational context it needs. Second, workflows: winning teams redesign entire end-to-end processes rather than simply accelerating isolated tasks. As Sands puts it, speeding up individuals who are pointed in slightly different directions causes them to quickly crash into each other. Third, culture: the fastest-moving teams work under leaders who explicitly encourage learning and experimentation, making it clear that some experiments will fail.
Sponsored Protocol
Overcoming the hidden knowledge barrier
Sands argues that another obstacle is not technology itself but employees figuring out AI on their own. Every worker develops different prompts, agents, and assumptions, creating a layer of unspoken knowledge that rarely translates into organizational performance. To counter this, Atlassian experimented with AI working agreements at the start of projects: teams decide not only what they will use AI for, but what they will deliberately avoid, which agents they will share, and what common skills will keep everyone working from the same context. Teams that adopted this practice used AI more, moved faster, made better decisions, and produced higher-quality work.
Sponsored Protocol
Experimentation and constraints as learning accelerators
According to Sands, the fastest route to learning is a combination of experimentation and constraints. Teams seeing the biggest gains deliberately imposed constraints, such as breaking every task into the smallest practical unit of work (a single story point) or committing to write no code by hand for a week. Though not sustainable forever, this approach accelerates learning. The broader lesson is that AI does not create entirely new management problems; it exposes old ones. Teams have always struggled with hidden assumptions and different mental models, and AI makes those gaps more consequential, increasing the importance of shared context and explicit ways of working.
Sponsored Protocol
For more on the challenges of AI adoption, read the related article on OpenAI builds an autonomous hacker. More about Atlassian can be found on Wikipedia.
