APPLICATIONS
“ Your agent can create way more queries in parallel than you can,” notes Markus.“ Bandwidth requirements go up a lot and latency becomes more important as agents talk to each other.”
Network traffic driven by agents is also far less predictable than traffic driven by humans.
“ If you send an agent off, it spawns over 100 different jobs that you would never be able to do yourself,” Markus continues.“ The network traffic patterns will look very different – very spiky, very high performance, but low latency and reliability are going to become critical.”
Markus illustrates the scale of the change by explaining that wireless networks are typically designed around an estimated number of human users per access point, based on how much bandwidth each person needs.
“ If you run 10 agents in addition to yourself on one device, you suddenly have 11 users from that single device,” he says.“ The density of users – humans and agents combined – is exponentially increasing, and so are the bandwidth requirements.”
Wireless networks, not to mention the layers connecting them to the internet, need re-architecting to handle that surge, Markus adds.
Ensuring AI readiness For companies yet to deploy AI agents at scale, Markus recommends starting with an honest audit. That means assessing current network capacity alongside every AI project already under way, partly for governance reasons.
AGENT ONE COWORKER TURNS INSIGHT INTO ACTION
Extreme Networks recently unveiled Extreme Agent ONE Coworker, which it says can resolve IT issues up to 15 times faster. The new AI agent combines realtime network context, historical trend analysis and agentic reasoning to deliver actionable recommendations into daily workflows – providing IT teams with the precise context they need so they can spend time solving problems rather than searching for them.
“ Extreme Agent ONE Coworker understands the intricate details of your network environment, encompassing historical client experience and network performance,” states Nabil Bukhari, CTO and President of AI Platforms at Extreme Networks.
“ That context is the difference between troubleshooting a ticket and fixing the root cause of a problem with documents analysis and recommendations. It turns insight into action and problems into resolutions at machine speed, while keeping people firmly in control.”
102 October 2026