Introducing Kentik AI Advisor: The Future of Network Intelligence
Summary
Introducing Kentik AI Advisor, a powerful new AI designed to deeply understand your network, reason through complex issues, and deliver clear, actionable guidance for designing, operating, and protecting your networks. By autonomously querying Kentik’s rich telemetry and tools, it explains what’s happening, why it matters, and what to do next — from troubleshooting and capacity planning to cost optimization and risk mitigation.
It’s almost cliché to say that change, complexity, and scale in IT are accelerating — but it’s undeniably true. It feels like only yesterday companies were racing to the cloud as the next big transformation. Now, just a few years later, AI is reshaping the fabric of society at an even faster pace. Whatever the future holds, one thing hasn’t changed: The network remains the critical foundation everything depends on.
The problem is that networks are also subject to this rising complexity, yet teams have the same budgets and fewer hours to keep it all working. At Kentik, we’ve always believed that data without context is just noise — and that true network intelligence requires the ability to understand what’s happening, why it matters, and what to do next.
But as networks stretch across clouds, fabrics, edges, and global backbones, achieving that intelligence has become harder. Outages at major companies remind us that the gap between complexity and human capacity is getting wider.
Today, that changes with the introduction of Kentik AI Advisor — the first AI designed to think and work like network engineers do. It gives teams the speed, clarity, and intelligence needed to design, operate, and protect networks at a modern scale. It’s a new chapter for Kentik and, we believe, a turning point for how network operations will be done in the years ahead.
Meet Kentik AI Advisor
Kentik AI Advisor deeply understands your network, thinks critically, and advises how to design, operate, and protect network infrastructure at scale.
AI Advisor acts like an expert network engineer who knows your environment inside and out and is always available to help. It dramatically increases the speed at which teams can operate, empowers every engineer to perform like an expert, and helps organizations with limited resources gain control over ever-increasing complexity and scale. As a result, teams can regain time previously spent fixing or maintaining networks and reinvest it in strategic projects.
AI Advisor works on four key principles:
- Thinks like an engineer – AI Advisor interprets natural language requests and reasons through complex, multi-step investigations, just like an engineer.
- Acts independently – AI Advisor independently runs queries across all your network telemetry to solve the multi-step investigative plans it came up with according to your needs.
- Provides transparency and validity – AI advisor provides step-by-step logic and transparent data to validate how it reached its conclusions.
- Delivers expert advice – AI Advisor provides clear, concise, expert advice and actionable recommendations based on your network telemetry.
How does it work?
At its core, Kentik AI Advisor is powered by sophisticated context engineering, frontier reasoning LLMs, and an agentic architecture that leverages Kentik’s highly enriched network telemetry and tools.
The user describes what they’re trying to understand or accomplish.
Kentik AI Advisor applies contextual, domain-aware prompt engineering and builds a structured, enriched request to the LLM. The LLM determines which Kentik tools and data types are best suited to answer the question and issues tool-use requests back to Kentik.
The LLM and AI Advisor work together to query Kentik data and knowledge services and refine results. AI Advisor returns to Step 2 as needed.
Kentik AI Advisor returns a clear, evidence based explanation and recommendation tailored to the user’s original goal.
When you ask a question – say, “Why did latency spike in Frankfurt?” AI Advisor doesn’t just pull up metrics. It interprets your intent, then, based on its context and awareness of Kentik’s data, it determines what information is necessary to answer the question and what data it can access in Kentik to confidently provide an answer.
Once the plan is set, AI Advisor takes action. Through a recursive, agentic loop, it autonomously calls Kentik’s own tools — Data Explorer, NMS Metrics, Cloud Pathfinder, Synthetics, and more — to gather the evidence it needs. Importantly, it shows you its thinking at each step of the way and also creates data outputs so you can view and validate the data it references in its analysis.
It grounds every step of its reasoning in real telemetry: flows, traps, syslogs, cloud paths, and performance data from your environment.
Natural Language Runbooks and Custom Network Context
Every network team has its own way of doing things, its own playbooks for how to handle various situations. For those situations, Kentik AI Advisor goes beyond foundational LLMs and allows you to customize AI Advisor to your environment, team, or business needs.
The first way we do this is by aligning Natural Language Runbooks with alert policies to increase precision, specificity, reliability, and repeatability for specific alert troubleshooting scenarios. These are like the runbooks you might be familiar with from other tools, but better. Instead of rigid syntaxes or code-based runbooks, where every step has to be defined in YAML, JSON, or a visual, low-code flow builder, you write Kentik runbooks in natural language, the same way you’d explain the process to a colleague. AI Advisor turns those instructions into structured investigations and executes them automatically.
Natural Language Runbooks are a simple, yet powerful tool for empowering your people to act with precision, reliability, and expertise when time matters most.
The second way you can tailor AI Advisor to your network is through Custom Network Context. Where Natural Language Runbooks are specific to the alert policies you assign them to, Custom Network Context applies to any task you assign to AI Advisor. It is how you can apply specific, relevant institutional knowledge to AI Advisor, so it knows your network like you do. It’s how AI Advisor will know that when you ask why Steve isn’t responding, that you aren’t talking about Steve in HR, but Steve, the name you gave to your skunkworks server. Other great uses for this capability include adding your IP ranges, site names, service tiers, and maintenance windows. This custom network context lets it reason in ways that are specific to your environment, not generic best guesses.
Faster answers, better decisions, and scalable outcomes
Now that we’ve talked about how AI Advisor works, it’s worth focusing on why that matters: it helps teams deliver faster, more reliable outcomes across the work that matters most – troubleshooting, capacity planning, cost optimization, performance tuning, risk mitigation, etc. – even as the network keeps getting more complex.
AI Advisor can do a lot. During early access, customers have already discovered creative ways to use it – from rapid incident triage to capacity planning, DDoS investigation, and beyond.
Put simply, AI Advisor knows how to use all the same tools and telemetry you rely on today – flow, routing, device metrics, synthetics, cloud paths, logs, runbooks, and your own custom context – just faster, more consistently, and with deeper reasoning. If you were to use a combination of these to solve a problem, you can expect AI Advisor to do the same.
| AI Advisor |
|
Core / Flow
NMS
Cloud
Synthetics
Knowledge and Context
|
Take a simple example: troubleshooting a dropped BGP session. Normally, troubleshooting this type of incident means bouncing between tools, validating metrics, digging through logs, correlating flow changes, and relying heavily on experience to sort signal from noise. AI Advisor does all of this for you. It interprets the problem, runs the right queries, gathers the evidence, explains what’s happening and why, and delivers clear next steps – in a fraction of the time.
But the impact goes well beyond break/fix scenarios. Because AI Advisor can reason across all of Kentik’s telemetry – flow, routing, devices, synthetics, cloud paths, and your own runbooks and custom context – it helps teams in many powerful ways. Among them (but certainly not limited to):
Plan with confidence
Capacity planning has always been equal parts art and science, always aiming for that optimal balance of investment and capacity overhead to keep on top of future scale and growth. Getting that mix right involved a lot of manual data analysis to understand trends, identify potential bottlenecks, and determine the most cost-efficient options. AI Advisor not only helps you pull that data faster, it also helps you understand it and provides advice on what to do, so you can run through planning exercises faster, with more accuracy and insight.
Optimize cost and performance
Cost and performance optimization has traditionally meant long hours of digging through flow data, cloud bills, and routing tables to pinpoint where money or performance is not optimized. It’s slow, nuanced work that depends heavily on experience to know where to look and what matters, especially as more teams are required to work across cloud and traditional networks. AI Advisor takes on that heavy lift, helping you surface inefficient paths, unnecessary spend, such as overuse of cloud transit gateways, or suboptimal peering decisions. The result is faster, clearer decisions that help your budgets go further while making your network run smoothly.
Protect network reliability and resilience
When suspect traffic surges or a service degrades, figuring out whether you’re looking at malicious behavior, a misconfiguration, or just normal volatility can be time-consuming and stressful. AI Advisor helps by pulling the right telemetry automatically and analyzing it in context, so it can quickly separate legitimate from malicious traffic and provide an educated opinion and guidance on next steps.
Getting the most from AI Advisor
AI Advisor is at its best when it can see the full context of your network. If you only gather flow from a device but don’t have NMS to gather metrics, AI Advisor won’t be able to help answer questions that rely on both. That’s why AI Advisor is most powerful when paired with all of Kentik’s other products – Flow, NMS, Cloud, and Synthetics – so it can truly be a force multiplier for accelerating network intelligence.
AI Advisor is also designed so you stay in control. It doesn’t automatically push changes or implement actions on its own. It surfaces options, explains trade-offs, and recommends next steps – but you decide what to do and when. It is also worth noting that Kentik AI Advisor exchanges data with LLMs only for inference and is not used for training purposes. Organizations also have the flexibility to use their preferred LLM provider via their own key or choose to opt out of Kentik AI entirely.
Over time, we’ll continue expanding what data AI Advisor can access and what it can do with that data; we’ll be taking an AI Advisor-first approach to new features and capabilities in Kentik moving forward.
The future of network intelligence
AI Advisor marks the beginning of a new era for network intelligence, where scale and complexity don’t bog down teams.
Today, AI Advisor can reason, analyze, and advise with the full context of your network. Tomorrow, it will go even further, proactively identifying issues before they become incidents, running analyses on its own, and automatically recommending actions. It will spot when traffic patterns shift in ways that signal congestion, when a configuration change creates issues in the network (stay tuned for this exciting announcement in the new year!), or when capacity is about to run short.
With AI Advisor, Kentik is building toward a future where network intelligence is always on and the continuous work of designing, operating, and protecting your network isn’t a balancing act of competing priorities, but instead becomes strategic, rewarding, and impactful for your business. We are building a future where teams move faster, make better decisions, and stay ahead of growing complexity even when budgets and headcount don’t. That’s the promise of network intelligence, and we’re just getting started.
Want to see AI Advisor in action? Request a demo today.
