Enterprise AI agent adoption jumps 466.7% as companies expand automation
The rapid growth of AI agents is prompting organizations to rethink how they manage access, permissions and risk across digital environments.
BeyondTrust reported that enterprise AI agent adoption increased 466.7% year over year and outlined five recommendations to strengthen governance and security.
The adoption of artificial intelligence agents within organizations continues to accelerate as businesses seek to automate tasks and improve productivity.
According to research conducted by BeyondTrust Phantom Labs, AI agents operating in enterprise environments recorded year-over-year growth of 466.7%. As these tools become embedded in daily operations, they are increasingly interacting with corporate applications, critical systems and sensitive data.
BeyondTrust warned that enterprise AI agent adoption rose 466.7% year over year and proposed five recommendations to strengthen governance and security.
Five recommended measures.
The company said the challenge for organizations is no longer limited to adopting artificial intelligence, but also includes securely managing the identities and permissions associated with AI agents.
Many of these systems require access to the internet, enterprise tools, platforms and endpoints in order to perform autonomous actions, which can increase risk when they operate with excessive privileges or insufficient controls.
According to BeyondTrust, a lack of visibility and governance can turn AI agents into a potential attack surface. In response, the company outlined five measures to strengthen the secure management of AI agents in enterprise environments.
1) Treat AI agents as digital identities.
BeyondTrust noted that these systems can access corporate resources, move information and perform actions, and therefore should be managed in the same way as users or service accounts with access to enterprise systems.
2) Apply the principle of least privilege.
Each agent should be granted only the access required to perform a specific task. According to the company, this practice helps reduce risks associated with errors, misuse or unauthorized activity within corporate environments.
Visibility and credentials.
3) Maintain visibility into AI agent activity.
Organizations should understand where agents operate, which systems they use, what resources they can access and what actions they are authorized to perform in order to identify risks and insecure configurations.
4) Protect the credentials and secrets they use.
BeyondTrust explained that many agents rely on API keys, tokens, passwords or service credentials, and recommended that these be stored and managed securely to prevent exposure that could be exploited by attackers.
In practice, securely managing AI agents requires combining controls over identities, privileges, access and activities to reduce risk in increasingly automated environments.
5) Continuously monitor and review agent activity.
Because these systems can execute multiple actions within seconds and operate at scale, the company said organizations should log activities, detect unusual behavior and periodically review permissions to ensure they remain aligned with business requirements.
Governance and oversight.
To support this process, BeyondTrust advocates a least-privilege approach that grants only the access required for a specific task and only for the necessary duration.
The company said that managing identities, permissions and activities allows organizations to gain greater visibility into the risks associated with AI adoption, strengthen access governance and improve oversight of privileges in increasingly automated environments.
Taken together, BeyondTrust’s recommendations are intended to help organizations implement identity, access and monitoring controls to manage the growing use of artificial intelligence agents more securely across their operations.
|
|
