Nutanix’s Kanuj Behl on Moving AI From Pilots to Production

Nutanix’s Kanuj Behl on Moving AI From Pilots to Production

By
Publication Date: 2026-11-19 16:07:00

As organizations move artificial intelligence (AI) from experimentation into production, they face new questions about cost, security, access, and governance. Kanuj Behl, principal enterprise architect, Nutanix discusses how organizations can manage those challenges, why flexibility across models matters, and what he hopes to hear from peers at the Tech Tonic: GTC Keynote Watch Party.

MeriTalk: As organizations move from experimenting with AI to deploying autonomous agents, what new challenges are emerging around cost governance and control?

Kanuj Behl: Moving from a pilot to production is a natural evolution. During testing, a small group of people may be working with a minimally viable product. Once an organization puts it in the hands of more users and introduces more automation, it needs visibility and controls around access, governance, and cost.

Testing can allow more flexibility. Regular use is different. Organizations need to plan for how they will manage these systems at scale and keep them within real-world budgets.

MeriTalk: AI adoption is accelerating, but increased usage can also mean increased costs. How can organizations scale AI without letting token consumption drive unpredictable spending?

Behl: Start by tightening the prompts and instructions used in production. Then choose the right model for the task. A simple request may not need the most capable model available.

Organizations should also review which tools and servers they use, who has access, how…