Extreme launches Agent ONE Coworker for network teams

Extreme Networks has made Agent ONE Coworker available globally to customers of Extreme Platform ONE.

The software is included with Platform ONE subscriptions worldwide.

The launch marks Extreme's latest effort to use artificial intelligence to address a familiar problem in corporate IT operations: network teams often spend long periods gathering information from different systems before they can identify the cause of an outage or performance issue. The new product is designed to reduce that manual work by mapping relationships across users, devices, applications, services and network conditions.

Network context

Agent ONE Coworker sits within Extreme Platform ONE and is aimed at enterprise networking teams. According to Extreme, it provides answers based on the specific conditions of a customer's network rather than broad, generic models. It also shows its reasoning alongside recommendations so engineers can review the basis of a suggested action.

Extreme is positioning the product as a more proactive form of AI support for network operations. One new feature, called Nudge, analyses historical network baselines, current traffic patterns and behavioural trends across the platform to identify issues that may need attention before a user asks for help.

Other functions include Talk to RRM, which focuses on wireless radio performance, and a canvas tool that creates dashboards and reports for different audiences. The product also includes Talk to Knowledge and Talk to Data tools, which Extreme said have improved response accuracy through the use of its knowledge graph.

Extreme said the software can help customers resolve issues up to 15 times faster. It also said some functions can cut the time needed to onboard new team members by up to 50%, reflecting the wider pressure on IT departments to do more with limited specialist staff.

This approach comes as networking suppliers race to move beyond anomaly detection and chatbot-style interfaces towards systems that can interpret operational context, recommend actions and feed those suggestions directly into support workflows. In that market, vendors are trying to show that AI tools can shorten incident response times without removing human oversight from critical infrastructure decisions.

AI support

Nabil Bukhari, Chief Technology Officer and President of AI Platforms at Extreme, said the distinction lies in how much network-specific context the system can use.

"Generic AI broadly understands networking. Extreme Agent ONE Coworker understands the intricate details of your network environment, encompassing historical client experience and network performance. That context is the difference between troubleshooting a ticket and fixing the root cause of a problem with documented analysis and recommendations. It turns insight into action and problems into resolutions at machine speed, while keeping people firmly in control. It's not replacing expertise; it's scaling it across the enterprise," said Nabil Bukhari, Chief Technology Officer and President of AI Platforms, Extreme.

Extreme also said its Talk to Support function can use live network context to recommend fixes and, where necessary, escalate cases automatically to the company's technical assistance centre. That could reduce the back-and-forth typically required when support teams need to reconstruct the state of a network before advising on a remedy.

The company is also leaning on governance as part of its pitch to corporate buyers. It said usage tracking, logging, policy controls and audit trails are built into the platform, while final decisions remain with human operators. That emphasis reflects a broader concern among enterprise customers that AI systems used in infrastructure management must be explainable and controllable, particularly when recommendations may affect core communications systems.

Customer interest

Customer interest is already being framed around the ability to tailor responses to different levels of the organisation. Vandalia Health said the product could help both senior leadership and technical staff draw different kinds of insights from the same data.

"We're excited to get started with these new tools. Extreme Agent ONE Coworker will give us the ability to instantly turn network data into context-based insights tailored to each specific audience, so we can quickly get high-level answers that our CIO can use with leadership teams, while our IT team can dive deeper to troubleshoot and optimise network performance," said Cord C. Scott, Principal Network Engineer, Vandalia Health.

Industry analysts are also watching whether networking AI products can shift from assistants that answer queries to embedded systems that shape day-to-day operations. That change would put more weight on context, transparency and trust than on simple conversational features.

Industry outlook

"Enterprise networking is moving beyond first-generation AI. The next era is about context: AI that understands each customer's unique environment, explains its reasoning, and gives network teams confidence to act. Extreme Agent ONE Coworker is a strong example of that evolution, moving AI from a helpful assistant to a trusted part of network operations," said Shamus McGillicuddy, VP of Research, Enterprise Management Associates.