About This Episode
In this episode of Stratola Spectrum, Dinesh Chandrasekhar, Chief Analyst, Stratola, sits down with Bob Friday, Chief AI Officer at HPE, to explore how AI is reshaping enterprise networking and operations from the inside.
Bob’s career spans four decades of innovation. From founding Airespace and Mist Systems to leading AI strategy at Juniper and now HPE, he has been at the center of every major shift in networking architecture. His journey mirrors the industry’s own evolution, from hardware-centric systems to cloud-native platforms, and now to AI-driven automation.
This conversation moves beyond surface-level AI hype. Instead, it dives into how AI is being embedded into real production environments where uptime, resilience, and security are non-negotiable.
Bob explains the paradigm shift from managing network devices to managing user experience, measured minute by minute. He unpacks the distinction between traditional supervised AI models and today’s emerging agentic AI systems that operate as non-linear, reasoning-driven automation layers.
The discussion also covers:
- How cloud AIOps started long before the GenAI explosion
- Why self-driving networks are ultimately a trust problem, not a technology problem
- The convergence of networking telemetry and security intelligence
- The role of explainability in enterprise AI adoption
- Whether AI will augment or replace IT operations teams
If you are responsible for networking, AIOps, AI strategy, or enterprise infrastructure, this episode offers a grounded view of what AI looks like when it moves from research labs into mission-critical systems.
This is not theoretical AI. This is AI under operational pressure.
Key Takeaways
CAIO role is about turning network ops into “user experience ops.” Bob’s journey from wireless engineering to CAIO centers on shifting from managing devices to managing end user, client-to-cloud experience, using “user minute” as the core unit of measurement.
His 2025 priorities split into Horizon 1 vs Horizon 2. Horizon 1 is shipping near-term customer features: get the right telemetry to the cloud, normalize it into the user-minute model, and build explainable models. Horizon 2 is agentic AI that turns insights into self-driving, actionable automation.
Two AI tracks run in parallel: classic ML plus agentic reasoning. Classic supervised models predict experience (Zoom/Teams as the “canary”), and tools that explain which network signals drive poor performance. Agentic AI then reasons across many sources (docs, KB, tickets, bugs, telemetry) via MCP-style access and multi-agent routing.
AIOps is still about cloud first, then autonomy. The “get data to the cloud” step is the hard cultural change (breaking CLI and SSH habits). Once that foundation exists, agentic AI becomes the catalyst for moving from assistive insights to self-driving remediation.
Trust and explainability are the gating factors for self-driving networks. Bob compares it to Waymo adoption: teams gradually expand what they allow automation to touch after repeated success. Customers demand “how did you get that answer” before they let AI act on the data plane, and ROI shows up as less firefighting and more proactive ops time.