Scalable AI and token reduction with data curation. With Rick Fulwiler, NETSCOUT

AI promises higher efficiency, more automation and greater user experience. Yet it may also fail to scale or to deliver, or require more resources than it saves. How to get the good sides of AI and protect network operations from its challenges?

In this Sparring Partners, I talked with Rick Fulwiler, Senior Director Product Management at NETSCOUT, about how to make AI efficient through token reduction and how this is done primarily by curating the data to make sure the AI models and agents don’t get sidetracked or tempted to hallucinate.

Data curation is not just about selecting accurate data. It is about figuring out which data sources, and at what granularity, are relevant to the user's intent. In a way, this is an art steeped in human expertise, but it has to recognize that the data is used by AI rather than humans.

These are some of the highlights of our conversation:

• Why AI struggles with traditional network KPIs
• Packet data versus alarms and dashboards
• Data curation reduces cost and improves accuracy
• AI agents need governance and transparency

More from NETSCOUT

More on Senza Fili at https://senzafili.com

Sign up for the mailing list at https://senzafili.com/senza-fili-mailing-list/ to receive invitations to Sparring Partners events and updates on what we do.