Atlassian says AI is speeding up workers more than teams

Atlassian used a VB Transform 2026 fireside chat to argue that enterprise AI adoption is being aimed at the wrong unit of work: the individual employee rather than the team. No funding, acquisition, product launch or financial commitment was disclosed, but the claim is relevant to CIOs and operators trying to turn broad AI usage into measurable operating gains.

Dr. Molly Sands, head of Atlassian’s Teamwork Lab, told VentureBeat senior technology contributor Sam Witteveen that many companies are training workers to use AI tools without changing how teams coordinate work. Sands leads a group of behavioral scientists and psychologists that studies how AI affects collaboration and uses those findings to change work patterns inside organizations, according to Atlassian.

The gap Atlassian is pointing to is familiar: employees report faster task completion, while leadership struggles to connect that activity to company-level results. Sands cited Atlassian’s annual State of Teams Report, which surveyed 12,000 knowledge workers globally and included interviews with about 200 Fortune 1000 executives.

According to Sands, 89% of those executives said individual employees were moving faster because of AI. Only 6% said they could identify clear examples of return on investment. Atlassian also found that about 14% of teams had turned AI use into what it described as real value, though the company did not disclose a standardized financial threshold for that value in the discussion.

What Atlassian says separates higher-performing teams

Sands said the teams seeing better results had three common traits: context, workflows and culture. In Atlassian’s framing, context means putting goals, decisions and institutional knowledge into shared digital systems rather than leaving them in employees’ heads or scattered across messages.

Atlassian calls that shared layer a context graph. Across tools such as Jira and Confluence, the company says it connects work items, goals and the people responsible for them, giving AI systems more usable organizational context. That claim also aligns with Atlassian’s commercial interest in making its collaboration products the system of record for AI-enabled work.

Workflow design was the second factor. Sands said the stronger teams changed full processes rather than using AI to speed up disconnected tasks. Her argument is that individual acceleration can create more coordination problems if employees are optimizing local work while the overall process remains unchanged.

The third factor was culture. Sands said teams with leaders who encouraged experimentation and accepted failed tests learned faster. Examples included breaking work into the smallest practical units and, in some cases, asking software teams to spend a week writing no code by hand. Sands described such constraints as useful for learning, while acknowledging they are generally not sustainable as permanent practices.

AI working agreements

Sands also said companies are creating new pockets of hidden knowledge by letting each employee develop separate prompts, agents and assumptions. Atlassian’s proposed fix is an AI working agreement at the start of a project.

Under that approach, teams decide where they will use AI, where they will avoid it, which agents they will share and what common skills are required. Sands said teams using these agreements adopted AI more heavily, moved faster, made better decisions and produced higher-quality work. The company did not provide the underlying measurement detail for those outcomes in the session.

The broader message from Atlassian is that AI is exposing existing management weaknesses rather than replacing them. Teams already struggled with unclear decisions, fragmented context and mismatched assumptions. AI raises the cost of those gaps when employees can move faster in different directions.

This story draws on original reporting from VentureBeat.