Today, the Self-Driving Network™ is no longer a distant vision—it’s closer than ever to reality because of new agentic AI enhancements to Mist™, our AI-native networking platform. It’s how HPE Networking is accelerating the shift from “assisted” to “self-driving” network operations.
While others are making promises about agentic AI’s potential, we’re delivering on it today with a clear vision and proven results.
The agentic AI difference—and why a strong foundation matters
Agentic AI represents the next evolution in automation by introducing agents that can reason, plan, collaborate, and act across network domains—with little to no human intervention—to address issues proactively before they interrupt user experiences.
But for agentic AI to be truly valuable, it requires a strong foundation. At HPE Networking, we’ve been developing that foundation for over a decade with the Marvis® AI engine and Marvis® AI Assistant. The Marvis AI engine is continuously learning from relevant telemetry across networking domains and external applications—improving the efficacy of AI results, speeding up issue resolution, and assuring end user experiences from client to cloud. Our Marvis AI Assistant transforms how IT teams operate and interact with the network using industry-leading capabilities:
- Marvis Conversational Interface: Transforms how users interact with networks
- Marvis Minis: Proactively simulate user experiences to detect issues before they occur
- Marvis Large Experience Model (LEM): Provides predictive insights into network performance
- Marvis Actions: Enables automated remediation built on human-in-the-loop (HITL) trust models
With these core elements as the foundation, agentic AI is the catalyst to achieving self-driving networks faster. Today, we’re excited to announce new agentic enhancements that bring this vision closer to reality.
Enhanced Marvis AI Assistant: Agentic AI across all domains
Our latest enhancements to Marvis AI Assistant showcase agentic AI capabilities delivering real value across campus, branch, and data center environments. Our secure AI-native network extends agentic NetOps across operational domains, unifying operational teams to simplify client-to-cloud automated operations.
Enhanced Marvis Conversational Interface with multiagent collaboration
The enhanced Marvis Conversational Interface now uses GenAI to accelerate troubleshooting by leveraging agentic AI workflows and multiagent collaboration. Operations can now communicate with the network, asking open-ended questions such as, “Why is the Orlando site slow?”, and get intelligent analysis, contextual understanding, and precise resolutions in response. This evolution in Marvis AI Assistant marks a shift from assisted operations to autonomous networking intelligence, simplifying tasks like dashboard generation, cross-domain data correlation, and issue resolution through natural language input.
Expanded Marvis self-driving actions
Marvis Actions represents the evolution from reactive troubleshooting to proactive network operations. It identifies critical issues like DHCP failures, missing VLANs, and network loops, providing clear, evidence-based recommendations for rapid remediation.
Further, Marvis Actions offers automated remediations for IT-approved scenarios. Using an HITL trust model, teams can develop confidence over time, giving Marvis AI Assistant permission to self-drive, automatically resolving issues. Today, we’re announcing expanded self-driving actions, including:
- Correcting VLAN misconfigurations
- Shutting down ports to resolve network loops
- Upgrading noncompliant devices
- Handling routine policy updates and firmware compliance
- Resolving port-stuck issues and misconfigured ports
Each action, whether initiated by IT or executed autonomously by Marvis AI Assistant, is validated post-remediation and logged in the Marvis Actions Dashboard. This maintains full auditability and HITL oversight while building trust through consistent, accurate results. As IT teams witness Marvis AI Assistant successfully handle routine issues, they gain confidence to enable autonomous actions for increasingly complex scenarios. This not only speeds up time to issue resolution but also frees busy IT teams to focus less on repetitive tasks and more on high value, strategic work.
Generalized Large Experience Model (LEM) enhanced with Marvis Minis
Marvis Minis serves as an always-on digital experience twin that continuously monitors network conditions, capturing telemetry that fuels the Marvis LEM. Marvis LEM combines rich data with Shapley modeling to predict user experiences and pinpoint root causes before users notice issues. Through Shapley, Marvis AI Assistant provides intuitive visualizations that rank network elements by “guilt”—showing how much each contributes to a degraded user experience across APs, clients, and network features to quickly triage root cause.
Previously limited to Zoom® and Microsoft Teams® application data, Marvis LEM now leverages generalized training through Marvis Minis to deliver broader user experience insights—extending its value to customers beyond Zoom and Teams environments. These enhancements further reduce troubleshooting and ensure critical business applications are always optimized for peak performance.
AI for data center operations
We’re also thrilled to announce enhancements to Marvis AI Assistant that are specific to data center operations.
In data center environments, the Marvis Conversational Interface integrates with Apstra® Data Center Director’s contextual graph database, enabling the agentic AI framework to understand complex queries, break them into logical components, and iteratively query data sources to synthesize actionable responses. This framework currently supports nearly 300 API queries. It will expand to enable autonomous service provisioning activities, incorporate additional data sources like elastic search, and enhance feedback mechanisms for continuous learning—critical steps toward fully self-driving data centers.
We’re also extending Marvis Minis capabilities to data center operations, providing proactive intelligence that ensures the reliability of critical core services—such as DNS, network storage, and authentication services—from various devices and locations. Network operators can activate Minis for specific scenarios like post-maintenance validation or run them autonomously with configurable intervals to detect issues caused by network changes or failures, allowing customers to find “needle in a haystack” issues quickly and effectively.
The future of networking is self-driving
Our proven agentic AI capabilities, built on the strong foundation of our Mist platform, accelerate our Self-Driving Network vision from assisted operations to truly autonomous network solutions. This isn’t a distant vision. Our customers are operating self-driving networks today, with agentic AI agents that collaborate across domains, learn from every interaction, and deliver autonomous remediation that keeps users connected and productive.
Experience the new era of networking NOW
Ready to see how agentic AI can transform your network operations? Join our AI-Native NOW virtual event on September 16 or 17 to hear from HPE Networking leaders as they showcase how AIOps and agentic AI are driving automation that simplifies operations and elevates user experiences.
