Extreme Networks unveils AI agent saving 1,000 hours in troubleshooting, previews on-premises AI zones

Agent ONE Coworker delivers early results

Extreme Networks introduced Agent ONE Coworker, an AI agent integrated into its Platform ONE ecosystem, which saved customers approximately 1,000 hours of troubleshooting in its first month of deployment. According to the company, 41% of Platform ONE customers adopted the tool for daily tasks, with users responding to 98% of its critical alerts. Engagement with the AI agent was reported to be ten times higher than Extreme’s first-generation conversational assistant, signaling a shift in how AI-driven networking tools are being adopted.

The tool is designed to augment network teams by automating routine tasks and providing proactive nudges, rather than replacing human oversight. Nabil Bukhari, Chief Technology Officer and President of AI Platforms at Extreme Networks, emphasized that the goal is to give small teams the reach of larger ones without displacing existing roles.

Extreme Exchange zones leveraging network infrastructure

Extreme Networks previewed Extreme Exchange zones, a new initiative set to launch in the first half of 2027. The zones will repurpose spare compute capacity from access points and switches to support on-premises AI applications. This approach aims to address the challenge of discarded network telemetry, which often exceeds operational needs but is discarded due to cost constraints.

The initiative targets business use cases such as healthcare and sports fan experiences. For example, a fall-detection application demonstrated at the summit ran entirely on access points, while another use case for the Seattle Seahawks predicted gate congestion and identified fan walk times to pro shops. Extreme Exchange zones will operate on a consumption-based pricing model, measured in “agent actions.”

Governance and autonomous AI tiers

Extreme Networks also provided a preview of Agent ONE Operator, an autonomous AI tier slated for release in the first half of 2027. The Operator mode includes 268 tools, each configurable to auto-authorize, ask for permission, or block entirely. The company has established four baseline rules for the AI agent: do no harm, never expose credentials, never act outside its role, and never operate in another tenant’s environment.

Bukhari emphasized that governance for AI agents must reside outside the system running the agent to prevent rogue behavior. This approach aligns with the company’s broader strategy of embedding AI governance into its infrastructure, ensuring compliance and security for enterprise deployments.

Customer and industry reactions

Customers such as Hill College, the Institute of Cancer Research, and Liverpool Football Club shared positive experiences with Extreme’s AI solutions at the summit. Kevin Rocha, Chief Information Officer at Hill College, highlighted the tool’s impact on operational efficiency, while Jonathan Monk of the Institute of Cancer Research noted improvements in network reliability for critical research environments.

Andrew Grill, a keynote speaker and author at Digitally Curious, described ambient AI as a paradigm shift, outperforming conversational AI by proactively prompting users instead of waiting for queries. Industry analysts, including Zeus Kerravala of ZK Research, praised the potential of Extreme’s approach to sovereign AI infrastructure.

AI infrastructure as a competitive edge

Extreme’s focus on AI-driven network infrastructure follows broader industry trends, such as IBM and CoreWeave’s collaboration to develop controls for AI agent workloads. These efforts address challenges like workload isolation, identity management, and security, which are critical for enterprises deploying AI at scale. Extreme’s on-premises approach, using existing network hardware, could offer a cost-effective alternative to cloud-based AI solutions for businesses with stringent data governance requirements.

Similarly, TOTVS’s expansion of its enterprise AI foundation, LYNN, demonstrates the growing demand for AI solutions that integrate operational data with infrastructure. Extreme’s Exchange zones align with this trend, offering a scalable way to leverage untapped compute resources within enterprise networks.

What this means

Lazyfounder analysis — our interpretation, not reported fact.

Extreme Networks’ announcements reflect a broader industry shift toward integrating AI into foundational infrastructure, rather than treating it as a standalone tool. By embedding AI into network devices, the company is positioning itself to capture demand for cost-effective, on-premises AI solutions—particularly in sectors where data sovereignty and real-time processing are critical.

For founders and operators, Extreme’s model offers two key insights: first, existing hardware (like access points and switches) can be repurposed for AI workloads, reducing the need for costly new infrastructure. Second, the success of Agent ONE Coworker demonstrates that AI agents can drive measurable efficiency gains in network management, a use case that could extend to other domains like IT operations or customer support.

The focus on governance and autonomous AI tiers also underscores the importance of building trust in enterprise AI. Startups looking to compete in this space will need to address compliance, security, and explainability from the outset.

Key takeaways

• Agent ONE Coworker’s early success signals growing adoption of AI-driven tools for network management, particularly for automating troubleshooting and operational tasks. Founders and operators should assess whether similar AI agents could address inefficiencies in their own IT or customer support workflows.
• Extreme Exchange zones represent a novel approach to on-premises AI, repurposing existing network hardware for business use cases. This could reduce costs and improve data sovereignty for enterprises, especially in regulated industries like healthcare and finance.
• The consumption-based pricing model for Exchange zones (“agent actions”) introduces a flexible way to pay for AI services. Operators should evaluate whether this model aligns with their usage patterns and budget constraints when considering similar deployments.
• Governance frameworks, such as Extreme’s baseline rules for AI agents, are becoming essential for enterprise AI adoption. Startups building AI-driven products must prioritize compliance and security to meet enterprise requirements.

Related on Lazyfounder

• IBM and CoreWeave co-design workload controls for AI agent infrastructure
• TOTVS expands enterprise AI foundation LYNN with operational data integration
• More technology news