Tackling agent sprawl across the enterprise
Atlassian introduces governance tools to curb AI agent sprawl, offering visibility and controls to prevent unmanaged automation risks while enabling safe scaling across enterprises.
AI agents deployed without shared oversight risk creating ‘agent sprawl,’ where unmanaged tools operate with sanctioned credentials yet evade detection, according to Atlassian. Gartner projects Fortune 500 enterprises will average over 150,000 agents by 2028, yet only 13% believe they have adequate AI governance in place. Unlike shadow IT, agent sprawl involves unapproved actors using company data and permissions at scale, obscuring ownership, data access, and actions taken. Without visibility, leadership faces heightened complexity, technical debt, and security risks, as 94% of organizations report concerns over AI sprawl increasing these issues.
Atlassian’s solution centers on layered governance, starting with defining agent autonomy levels to balance automation with human oversight. Admins can restrict which actions agents may perform, on which data, and for whom, while teams retain flexibility to experiment within set boundaries. Agents surface only content the invoking user is permitted to access, and role-scoped controls apply agent-specific permissions beyond inherited user rights. Per-agent configurations further narrow data sources, tool invocations, and permission boundaries, ensuring accountability for agent actions.
Visibility is critical to managing agent sprawl, and Atlassian’s Agent Insights provides administrators with real-time data on agent counts, invocation frequency, and usage patterns. This enables targeted cleanup, consolidation, or scope reduction of redundant or high-risk agents. The platform also logs agent contributions in tools like Confluence, distinguishing agent activity from human actions to maintain clear audit trails. These features aim to align agent deployment with organizational priorities while mitigating operational and security risks.
Despite growing AI adoption, only 12% of knowledge workers in a recent Pulse survey say their AI understands company context better than a junior employee, highlighting gaps in agent readiness. Atlassian cites research indicating teams with optimized foundations are 7.3 times more likely to report higher AI ROI, underscoring the need for structured governance and contextual alignment before scaling agent use.