SAN FRANCISCO, June 2 (Reuters) - Cisco Systems on Tuesday announced a new suite of software tools that businesses can use to build their own armies of bots known as AI agents, to protect their IT infrastructure against cybersecurity threats.
Cisco's announcement comes as Anthropic is set to release its Mythos model in the coming weeks, an AI tool that some experts fear could be used by hackers to turbo-charge cyber attacks.
Cisco Cloud Control, as the company calls the system, is designed to allow businesses and governments tocreate and manage AI agents that will watch over their systems and block and remove hackers, among other functions. Because cybersecurity threats are increasingly coming from swarms of AI agents acting on behalf of human hackers, IT managers need to respond in kind, Cisco executives told Reuters.
"You can no longer do things at human scale," DJ Sampath, senior vice president and general manager of AI software and platform at Cisco, told Reuters. "It has to be machine scale, from an operational perspective."
To build those AI defender agents quickly, Cisco believesits customers will want to use AI coding tools. The company also rolled out an app-store-like marketplace where companies can choose among coding tools, with the first offered being OpenAI's Codex, which will be embedded directly in the Cloud Control platform. Sampath said that Cisco will take a cut of the sales through the platform but has not yet determined the exact amount.
"We're working through the economics. You should expect that we will have some economics that favor us, because it costs non-trivial amounts to be able to leverage all of these pieces," Sampath said.
The Cloud Control software is available in North America on Tuesday, and the marketplace for third-party tools will come in the second half of 2026, Sampath said.
(Reporting by Stephen Nellis in San Francisco; Editing by David Gregorio)
