Cisco Cloud Control unites AgenticOps for IT infrastructure
Cisco's control plane for AI agent-based infrastructure management marks a significant convergence of previously disparate tools, according to analysts.
LAS VEGAS --– Cisco is intent on making each of its network devices an enforcement point for centralized policy that controls AI agents that manage infrastructure, or AgenticOps.
In pursuit of that goal, the networking giant launched a new umbrella software framework called Cisco Cloud Control in controlled availability here at Cisco Live this week. The product combines an AI assistant, telemetry dashboards, third-party agent management, MCP tools and API support for AgenticOps. The interface for Cisco Cloud Control will be AI Canvas, a collaborative workspace previewed last year. Later in 2026, the product will add Cloud Control Studio, which will include an Agent Builder and a low-code App Builder with OpenAI's Codex assistant built in.
Cisco Cloud Control, which replaces the previous Cisco Security Cloud Control, provides a common orchestration layer for observability and network management in addition to security. All are connected to cross-domain log data in the Splunk-based Cisco Data Fabric, launched last year and set to reach general availability over the next two months.
Cisco is telling a "more cohesive story" than ever, delivering on previously disclosed roadmap plans consistently over the last 18 months, said Mike Leone, an analyst at Moor Insights & Strategy.
"For a long time, people would poke at Cisco for the multiple management interfaces across all the products in the portfolio," Leone said. "Cloud Control is, I would argue, the most coherent platform argument they've made in years."
Cisco's AI Canvas will become the common interface for AgenticOps automation across infrastructure domains using the new Cloud Control framework.
The challenge: standing out from the crowd
Virtually every enterprise tech vendor, whether it specializes in infrastructure or not, has rolled out its own AI agent orchestration platform over the past year, complete with pledges of compatibility with third-party agents. Cisco Cloud Control will face off against other large IT infrastructure vendors, including Microsoft, Google, AWS, Dell and IBM/Red Hat, each of which also has a broad base of existing enterprise customers -- many of whom overlap with those of the others.
"Every single vendor has this challenge," Leone said. "If you were to separate the existing customers versus net new customers [for AI agent orchestration platforms], existing customers would absolutely [outweigh] net new customers."
Cisco has quickly assembled a broad list of more than 50 partners for cross-domain AI agent support, which is new for the company, but not new in the market, according to Jim Frey, an analyst at Omdia, a division of Informa TechTarget.
"It's bringing Cisco into what has already been a thriving partner environment for everybody else," Frey said. "It's good that folks who are dedicated to working with Cisco equipment will now have access to the same kind of capabilities and ecosystem that other vendors have already pulled off."
Read agentic data egress fine print
Potential data egress costs enterprises might incur when sharing data across agents is just one of the hot topics in the industry surrounding pricing for agentic tools. Some vendors have begun offering outcome-based pricing, while many charge by licensing tier, which include pools of AI tokens.
Cisco Cloud Control fits into the latter category: it will be priced in three tiers: Essentials, Advantage or Premium. Customers can also buy token packs a la carte.
Enterprises should press vendors for deeper pricing details, said Mike Leone, an analyst at Moor Insights & Strategy.
"Where does that meter sit when a non-Cisco agent calls into Cisco data?" he said. "And is that on you or is that on the other platform that the agent's coming in from? That's where the open ecosystem promise is going to get tested really fast.
"Tier creep is a big thing, too," Leone added. "All it takes is somebody who's asking way too many questions or asking it to do too many things. And does the Essentials tier really give you enough to run Cloud Control to its fullest extent?"
The network is the differentiator?
Networking is cool again. As we move to these highly distributed environments, the network becomes a key enabler for operationalizing AI.
Bob Laliberte,Analyst, TheCube Research
Among Cisco's differentiators are its purpose-built large language models,includinga frontier model and specialized models for network management and cybersecurity. And it has a long track record with network visibility and control mechanisms, which are crucial to multi-agent communication and governance, said Bob Laliberte, an analyst at TheCube Research.
"Networking is cool again," Laliberte said. "As we move to these highly distributed environments, the network becomes a key enabler for operationalizing AI. Even Nvidia is a very large networking company right now."
However, Cisco must compete in AgenticOps against other networking vendors that also offer AgenticOps features, including Arista and HPE, Frey said. Network-specific software vendors can also claim an advantage over switch vendors in AgenticOps for multi-vendor networking environments.
As a general observability and log analytics company, Splunk gives Cisco some credibility in multi-vendor management as well, as long as Cisco continues to run Splunk somewhat independently, Frey said.
"There's a honeymoon period when a company acquires another company, and it has a certain amount of credibility as a third-party supplier that erodes over time, because people understand where the investments will be made and where they will not be made," he said.
Network Actions to include digital twins
Cisco's new Network Actions features include a digital twin that enables users to experiment with AgenticOps without affecting production.
A new generation of Cisco switches and routers will support network-specific AgenticOps features called Network Actions, due out in beta this month. The update includes some typical features of AgenticOps tools, such as telemetry dashboards and root-cause analysis, but it also offers an alpha-stage digital twin feature that enables users to experiment with infrastructure changes without affecting production.
While trust appears to be growing in AgenticOps products, it's primarily being used so far in compute systems rather than networking, according to recent market research Frey has reviewed.
"Compute systems are cattle -- servers and containers and virtual machines are all designed to be disposable," he said. "But networks are still like pets; something goes wrong with a network switch, okay, now you've got a problem that's got a much broader impact, potentially."
Features baked into Network Actions and Cisco Cloud Control, including confidence and risk scores and digital twins, don't address all of the trust issues with AI agents for infrastructure automation, but do address some of the common early barriers, Leone said.
"Cisco's design is really solid here," he said. "Surfacing confidence and risk scores that are based on every agent and the digital twin component will all make 'human in the loop' involvement really fast."
Beth Pariseau, senior news writer for Informa TechTarget, is an award-winning veteran of IT journalism. Have a tip? Email her or connect on LinkedIn.
Dig Deeper on Systems automation and orchestration
