Think like an architect with Scott Robohn

As AI tooling in the network gets richer and more agentic, what’s the actual job of the human sitting behind it?

In this session, “Think Like an Architect,” Scott Robohn — CEO @ Solutional and Co-founder of the Network Automation Forum — makes the case that the real shift AI demands isn’t learning new tools, it’s elevating how network operators think: from executing tasks to designing workflows and systems.

He covers:

• The difference between using AI for toil reduction versus using AI as a thought partner for broader systems thinking
• Why critical thinking, not coding skill, is the differentiator as agentic tooling takes on more of the keystrokes
• How to blend deterministic tools (scripts, regression) with non-deterministic ones (LLMs, agents) depending on the problem
• Practical techniques to try today — letting an LLM interview you, using multiple models to avoid single-threaded thinking, and treating “AI slop” as squarely the reviewer’s responsibility
• Why vibe coding is great for prototyping but hits real limits before it belongs in production
• What “architect thinking” looks like when scoping what an agent can read, write, and access

The session is a call to move from asking “can I automate this task?” to “can I design a reliable workflow for this outcome?” — with the human firmly in the loop as the ruthless reviewer and systems thinker.