Most enterprises aren’t ready for the security challenge that agentic AI presents

At the beginning of last year, Cisco predicted that the trajectory of AI would move towards agentic solutions and platforms, and the company was spot on, as nearly every organisation is now prioritising the implementation of agentic AI within its environment.

While companies are moving as quickly as possible to be agentic AI ready, there is one aspect that has seemingly been placed at the back of their minds – security readiness.

In its latest insights, Cisco noted that a significantly high 85 percent of enterprises are already experimenting with agentic AI, but crucially, only 5 percent have stated that their organisations have the infrastructure in place to securely manage them.

“AI agents aren’t just making existing work faster; they’re a new workforce of co-workers that dramatically expand what organizations can accomplish. Projects shelved for lack of resources are now within reach. The only limit is imagination, and security teams are the key to unlocking this opportunity by making the agentic workforce safe enough to trust,” noted Jeetu Patel, president and chief product officer at Cisco, regarding the potential of agentic AI.

Mirroring the potential are the security risks that agentic AI holds, especially if organisations do nothing to change their state of readiness.

To that end, Cisco has outlined three areas that is addressing when it comes to agentic workforce.

“First: Protecting the world from agents, ensuring they can only act as intended. Second: Protecting agents from the world, ensuring they can’t be manipulated or corrupted. Third: Detecting and responding to AI incidents at machine speed and scale,” it explained in a release to Hypertext.

Moving forward, the company today is extending its Zero Trust Access to AI agents, which is designed to hold them accountable to a human employee and securing agentic actions. Some of the new capabilities being added to the mix include:

• Strict Access Control: Agents are assigned fine-grained permissions only for the specific tasks they perform or resources they need for a short duration, with all tool traffic routed through an MCP gateway to eliminate blind spots.
• Agent Identity Management: Customers can register agents in Duo IAM and map them to accountable human owners, ensuring every agent has a verified identity and enabling traceability of actions.
• Agent and Tool Visibility: Cisco Identity Intelligence discovers agentic and non-human identities to help organizations understand existing AI usage.

Added to this, Cisco has announced new solutions designed to better interrogate and test agentic AI before they are fully deployed. This is critical, according to the company, as real-world conditions are often quite different from what agents are tested on.

This disparity is often when problems arise, and therefore having a more robust way of testing is becoming increasingly important.

“To empower more organizations to meet this challenge head-on, Cisco is democratizing the industry-leading capabilities of AI Defense by launching Cisco AI Defense: Explorer Edition. This new self-service solution is built on the same core AI Defense Validation engine trusted by Global 2000 customers,” said Cisco.

“After signing up, users can begin red teaming the AI models and applications that will be deployed into agentic workflows to uncover susceptibility to attacks and measure risk posture before deployment. This toolkit enables AI developers, AppSec teams, and security researchers to build and secure AI agents,” it added.

Cisco is also leveraging the capabilities of Splunk when it comes to better handling the security challenges that agentic AI presents. Here, it has already moved to embed AI capabilities into key SOC workflows.

One of the areas this is designed to make an impact in is the expansion of Agentic SOC with specialized AI agents. The company highlighted that these include the Detection Builder Agent, Standard Operating Procedures (SOP) Agent, Triage Agent, Malware Threat Reversing Agent, Guided Response Agent, and Automation Builder Agent.

“By automating security workflows, security tasks shift from a bottleneck to an accelerator, enabling the SOC to move at machine speed and scale,” Cisco emphasised.

With agentic AI now top of mind for most organisations, security is an even greater imperative. With the data showing that only 5 percent are ready to handle the impact of agentic AI from a security perspective, many organisations will need to have more robust plans and solutions in place.

[Image – Photo by Philip Oroni on Unsplash]