ServiceNow has released an urgent security update to fix a critical vulnerability in its AI‑powered Now Platform after researchers discovered that threat actors were actively exploiting the flaw in targeted attacks. The vulnerability, tracked as CVE‑2025‑12420, allowed attackers to bypass key security controls and gain unauthorized access to sensitive data stored within enterprise ServiceNow environments.

According to ServiceNow and independent security researchers, the flaw stemmed from improper access validation within the platform’s AI‑driven automation components. By manipulating specific API requests, attackers could escalate privileges, access restricted records, and potentially pivot deeper into connected enterprise systems. The vulnerability was rated critical due to the widespread use of ServiceNow across government agencies, Fortune 500 companies, and global service providers.

CyberScoop reports that the flaw was exploited in a limited but sophisticated campaign targeting organizations in sectors including finance, healthcare, and technology. Attackers appeared to focus on environments heavily reliant on ServiceNow’s AI‑assisted workflows, which often handle sensitive operational data, ticketing information, and identity‑related records. While the company did not disclose the identity of the threat actors, researchers noted that the exploitation pattern resembled activity commonly associated with advanced persistent threat (APT) groups.

ServiceNow stated that it became aware of the issue after receiving reports from security partners and immediately began developing a patch. The company has now released updates for all supported versions of the Now Platform and is urging customers to apply them without delay. Organizations using older or customized deployments were advised to review access logs for suspicious API activity and enforce strict least‑privilege controls.

Security analysts warn that the incident highlights the growing risks associated with AI‑integrated enterprise platforms. As organizations increasingly rely on automated workflows and machine‑assisted decision systems, vulnerabilities in these components can create high‑impact attack paths that bypass traditional security layers.

ServiceNow emphasized that it has not observed widespread exploitation but continues to monitor for new activity. Customers are encouraged to enable continuous monitoring, review API permissions, and ensure that AI‑driven modules are properly segmented from critical systems.