Battle of the Bots: Which AI Model Actually Work for Network Engineering Tasks - Eric Chou
Battle of the Bots: Which AI Model Actually Work for Network Engineering Tasks
As AI models evolve at breakneck speed, network engineers face a critical challenge: how do we determine which AI tools actually work for our specific tasks? While researching my upcoming book on AI-assisted network automation, I discovered that simply asking ""which model is best?"" is the wrong question. Instead, we need a systematic evaluation framework to make informed decisions about when and how to deploy AI in our workflows.
This talk introduces a pragmatic approach to AI adoption in network engineering: building a decision tree that calculates the expected value of using AI for specific tasks (like generating MOPs, writing automation code, or analyzing configurations), combined with a structured AI evaluation system. I'll demonstrate three practical evaluation methods—expert reviews, automated code validation, and AI-assisted evaluation—that help you quantify the probability of success for different AI models on different tasks. By the end, you'll have a framework to answer not just ""which AI is best?"" but more importantly: ""should I use AI for this task, and if so, which one?”"
#networkautomation #artificialintelligence #aimodel #aimodels