As AI agents become embedded in everyday enterprise workflows, monday.com emphasizes the importance of granular permission management to ensure safety while enabling faster AI innovation.
- Least privilege principles reduce AI security risks from excessive permissions.
- Proper permission management avoids common human errors in access control.
- Incremental rollout with limited access helps AI teams move faster and safer.
What happened
monday.com, in its transition into an AI-powered work platform, has underscored permission management as a fundamental mechanism to keep AI agents safe as they operate within enterprise workflows. Ryan Nigro, the company’s security trust and enablement lead, explained that governance around permissions is not intended to slow down teams but to provide confidence in accelerating their AI work.
During the recent Elevate event, Nigro discussed how applying the cybersecurity principle of least privilege—traditionally used for human users—can be adapted for AI agents. This involves granting AI tools just enough access to perform their tasks without enabling all features or broad permissions at once, thereby minimizing security risks as AI capabilities expand.
Why it matters
As AI agents act on behalf of users with given permissions, any excessive or sprawling access can result in security vulnerabilities at machine speed. Companies without strict permission management risk accumulation of harmful privileges, especially as employees change roles over time, which can lead to significant breaches caused by human error tied to over-permissioning.
An IBM study cited by monday.com estimates that 80% to 90% of security breaches occur because of such human errors in access management. By limiting permissions to what is strictly necessary, companies can sharply reduce the attack surface and maintain control even as AI agents automate more complex tasks within workflows.
What to watch next
monday.com’s approach to AI deployment, focusing on incremental rollout with least privilege permission policies, serves as a blueprint for other enterprises aiming to embed AI safely into their operations. Observers should watch how this permission model scales as more AI agents integrate with business tools and handle sensitive data.
Additionally, the industry will be monitoring how AI governance frameworks evolve to incorporate dynamic and automated permission adjustments in response to changing roles and workflow requirements, ensuring ongoing security without impeding the pace of AI-driven innovation.