According to VentureBeat, Kentik just launched Kentik AI Advisor, which they’re billing as the first artificial intelligence that actually understands enterprise and service provider networks at a deep level. The timing couldn’t be more critical – about a quarter of U.S. network engineers are preparing to retire in coming years, creating a massive skills gap. Early users like Equinix’s Lucas Isidoro report tasks that used to take thirty minutes now take seconds. The AI leverages Kentik’s proprietary data engine that ingests a trillion telemetry points daily, combining LLM capabilities with network expertise. The global AI in networks market is projected to explode from $15.28 billion this year to approximately $192.42 billion by 2034, making this a potentially huge play.
How it actually works
Here’s the thing that makes this different from traditional AIOps tools. Most network AI just correlates alerts that are fed to it. Kentik’s approach is more like having an actual network engineer in the system – it can interpret what you’re trying to accomplish, build an investigation plan based on real network telemetry, and then execute that plan while explaining its logic. It’s pulling from flow data, device metrics, cloud information, and even internal runbooks to deliver tailored recommendations. Basically, instead of just showing you red alerts, it’s supposed to understand what “check why the East Coast users are experiencing latency” actually means and go figure it out.
The real network problem
Network teams are absolutely drowning right now. Between cloud complexity, hybrid infrastructure, and that looming retirement wave, there’s a perfect storm brewing. Traditional monitoring tools generate tons of data but not much actionable insight. You end up with engineers spending hours sifting through dashboards instead of actually fixing things. Kentik’s betting that by giving teams an AI that can do the investigative grunt work, they can keep networks running smoothly with fewer people. It’s a compelling pitch, honestly – who wouldn’t want to turn 30-minute troubleshooting sessions into 30-second conversations with an AI?
But will it deliver?
The promises are impressive – cost optimization through analyzing VPC and transit expenses, capacity planning that’s actually proactive, DDoS investigation that separates real threats from noise. But here’s my question: how much of this is genuinely new versus repackaged analytics with an AI chatbot frontend? Kentik’s been collecting massive amounts of network data for years, so they’ve got the foundation. The real test will be whether this can handle the weird, edge-case network issues that always seem to pop up at 2 AM. If it can actually understand network intent rather than just pattern-match, that’s a game changer. If it’s just a fancier query builder, well, we’ve seen those before.
The bigger picture
We’re seeing this across the infrastructure world – companies are desperate for automation that actually works. The talent shortage is real, and throwing more bodies at network problems isn’t a sustainable strategy. What’s interesting is Kentik’s focus on what they call “agentic AI” – systems that don’t just recommend actions but actually execute plans. That’s a step beyond most of what’s out there today. For industrial and manufacturing environments where network reliability is absolutely critical, having AI that can proactively manage infrastructure could be transformative. Companies like Kentik are pushing hard into this space, and honestly, the network operations world could use the help.
