The future of network operations has arrived, and it is agentic.
Gartner® predicts that “by 2030, 50% of organizations will use agentic NetOps with minimal human involvement, up from nearly....”
With Prisma SD-WAN, we recognize that traditional NetOps will go through the agentic transition. We see our customers moving away from a reactive "war room" culture into a modern, autonomous network operations mode. Prisma SD-WAN is at the forefront of this change, pioneering new ways for customers to interact with our system.
For network leaders, the stakes are high. Traditional Network Operations (NetOps) face limitations due to the expertise, time, scale, attention and criticality of issues handled by the NetOps team. Manually sifting through vast quantities of logs and correlating data from various system points, such as packet captures, flow logs and overlay/underlay connectivity, is both complex and time-consuming.
The multidimensional nature of network data further compounds this complexity. It is a complicated process to correlate static network policy configurations with dynamic data plane traffic.
In contrast, AI agents can execute the same troubleshooting procedures at scale, correlating dynamic and static data, significantly faster. Manual incident investigation can take hours, resulting in a high mean time to resolution (MTTR). Agents can monitor hundreds of data points such as interface status, logs, errors and recent configuration changes in near-real-time, effectively serving as a powerful force multiplier for the NetOps team.
NetOps teams are often risk-averse, focusing primarily on maintaining the status quo rather than driving efficiency improvements. A network blackout or application brownout typically forces teams into a reactive troubleshooting mode. An AI agent fundamentally shifts this dynamic, enabling a proactive investigation model.
For NetOps, the focus is shifting from automated playbooks, which handle static routine tasks, to AI agents that can adapt to fast-changing environments. AI agents can think, plan and execute actions autonomously, thus making them the ideal tool for dynamically changing complex network environments.
Prisma SD-WAN is spearheading this change by launching a specialized troubleshooting AI agent for NetOps. This agent is designed for the autonomous resolution of frequent networking problems, such as network blackout and application brownout situations. The system is designed as a hierarchy of agents that are experts on specific domains.
Prisma SD-WAN Troubleshooting Agent is a purpose-built AI agent trained on deep product mastery and core networking knowledge. This agent follows a structured, autonomous workflow as follows:
The agent immediately identifies the scope and severity of the incident.
Instead of a static script, the agent builds a custom troubleshooting plan based on the real-time context.
Extract real-time context for guided, simplified troubleshooting.
It automatically pulls telemetry, logs and configurations to validate its reasoning.
Correlate multidimensional data points for faster root cause analysis.
It delivers a precise Root Cause Analysis with a one-click option to fix the issue.
Identify root cause issues with recommended actions.
The Troubleshooting Agent provides enhanced capabilities, including:
These agents have been trained on internal knowledge base, remediation playbooks, best practice and configuration guides using retrieval-augmented generation (RAG) for better accuracy and context awareness.
Shows every step taken to reach a conclusion. Customers can click a link to see exactly how it reached its finding.
If the root cause is outside the SD-WAN (e.g., a switch port issue), the agent provides clear next steps, such as checking configurations on adjacent network devices and providing remediation recommendations.
With Troubleshooting Agent, identifying the deep root cause of a network blackout or application brownout and gaining immediate remediation steps is now near instantaneous, thus reducing the mean time to root cause.
Adopting an agentic model for network operations helps deliver clear, quantifiable improvements in organizational performance and efficiency.
Intelligent agents in Prisma SD-WAN provide data-supported root cause analysis (RCA) that pinpoints the exact reason for network issues.
These Prisma SD-WAN agents significantly improve their reasoning capabilities with RAG. RAG achieves this by integrating customer-specific data such as custom configurations, design documents, and knowledge base information, allowing the RCA to be precisely customized to the customer's unique network architecture, thus offering more targeted diagnostics than generic solutions.
The agentic network operations system employs a focused and highly precise Troubleshooting Agent designed for efficacy and reliability. This agent accelerates diagnostic cycles by concurrently running various diagnostic steps and tools across the entire SD-WAN fabric. To achieve specific goals and maintain clear task boundaries, the Troubleshooting Agent analyzes the issue and can leverage various backend tools, such as log and packet capture analyzers. Critically, the system incorporates human-in-the-loop oversight at all times and generates a clear remediation plan.
Minimizing network downtime directly translates to continuous productivity for branch users, sustained network uptime for mission-critical applications, and protection of revenue streams.
The autonomous era of NetOps is here. By granting your network the power to automatically identify and provide remediation steps, you not only streamline current operations but can also future-proof your business. Prisma SD-WAN enables this essential framework and AI agents required to begin your agentic NetOps journey. The future of NetOps is agentic. Are you ready?
To learn more about the AI innovations with Prisma SD-WAN read How Prisma SD-WAN AI transforms Network Operations: Strata Copilot.
Gartner, Agentic NetOps Will Revolutionize Network Operations, Jonathan Forest, Andrew Lerner, 06 August 2025.
GARTNER is a trademark of Gartner, Inc. and/or its affiliates.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
| Subject | Likes |
|---|---|
| 5 Likes | |
| 2 Likes | |
| 2 Likes | |
| 2 Likes | |
| 1 Like |
