Provided by Atlassian
During a fireside chat with VentureBeat Senior Technology Contributor Sam Witteveen at VB Transform 2026, Dr. Molly Sands, head of Atlassian’s Teamwork Lab, highlighted a common misstep in AI adoption among companies. Many organizations focus on optimizing individual AI usage instead of enhancing team collaboration.
Dr. Sands leads a talented team of behavioral scientists and psychologists dedicated to examining how AI transforms collaborative work dynamics. Their findings are instrumental in helping businesses redesign their operational processes.
“We don’t just conduct research; we actively implement changes,” she stated. As her team introduces innovative work methodologies and redefines workflow strategies, numerous organizations continue to wrestle with these complexities.
Understanding the AI ROI Gap
This year’s Atlassian State of Teams report, which surveyed 12,000 global knowledge workers and interviewed nearly 200 Fortune 1000 executives, revealed a stark disconnect between business activity and value generation. Despite widespread AI utilization, few can pinpoint its tangible benefits.
“While 89% of executives reported increased velocity among individuals in their companies, only 6% could provide clear examples of return on investment (ROI),” stated Sands.
Approximately 14% of teams are successfully translating AI use into measurable value. This indicates that while some organizations have high-performing teams, many others remain unprofitable.
Successful teams exhibited three essential traits: context, workflow, and culture. They established what Atlassian refers to as a context graph, integrating goals, decisions, and organizational knowledge into a shared digital framework rather than relying on individual memory. Across tools like Jira and Confluence, these graphs link tasks, objectives, and team members, offering the necessary context for effective AI utilization.
In terms of workflow, the most effective teams reimagined their entire end-to-end process instead of merely speeding up individual tasks. As Sands describes, failing to unify direction leads to inefficiencies, with team members “immediately starting to crash into each other.”
On the cultural side, teams that progressed quickly operated under leaders who promoted a culture of learning and experimentation, while also recognizing that not all experiments would succeed.
Empowering Leaders to Transform AI from Individual Efforts to Team Advantage
Sands emphasizes that experimentation within constraints is vital for learning. High-performing teams intentionally imposed restrictions on their work processes, breaking tasks down into manageable units (single story points), and even refraining from manual coding for a week.
“Most of these methods can’t be sustained indefinitely, but they facilitate rapid learning,” she remarked.
Another significant challenge is the varied approach employees take in navigating AI. Each employee tends to develop unique prompts, agents, and underlying assumptions, which creates a layer of tacit knowledge that often doesn’t translate into organizational performance.
To address this, Atlassian trialed AI working agreements at project kickoff, encouraging teams to clarify not only their AI usage but also their intentional exclusions, the agents to be shared, and the core competencies needed for a cohesive working context. Teams that embraced this practice utilized AI more effectively, acted more swiftly, made superior decisions, and ultimately delivered higher quality outcomes.
The overarching insight, Sands concluded, is that AI does not generate entirely new business challenges; it illuminates existing ones. Teams have long grappled with unspoken assumptions and divergent mental models. AI amplifies these discrepancies, underscoring the necessity for shared context and explicit working methods.
Sponsored articles are content created by companies that pay us to post or have a business relationship with VentureBeat, and are always clearly marked. For more information please contact us[email protected].
Source: venturebeat.com