Free Lesson

What Your AI Agents Actually Cost:The Hidden Economics of AI

Hosted by Vyoma Gajjar

Thu, Jul 23, 2026

7:00 PM UTC (45 minutes)

Virtual (Zoom)

Free to join

Invite your network

What you'll learn

Deconstruct multi-turn spend

Map how context regrowth, retrieval loops, and retries quadratically inflate bills.

Map cost-vs-quality curves

Use unit-test eval harnesses to find the exact threshold where saving money breaks agents.

Bound pre-production risk

Deploy simulation and edge-routing frameworks to lock down API exposure before going live.

Why this topic matters

AI models are getting cheaper, but enterprise AI bills keep growing. The reason isn't token prices: it's hidden agent behavior. Modern AI systems search, retry, call tools, and orchestrate multiple models before completing a task. In this live session, you'll calculate the true cost of a real AI workflow and leave with a practical framework and spreadsheet to audit your own systems.

You'll learn from

Vyoma Gajjar

AI Architect | Lecturer | Advisor | Scout

Hello, I'm Vyoma. I'm a Senior Principal AI Architect at ServiceNow, where I build the agentic AI systems and adoption programs that help large enterprises actually go live with AI, not just pilot it.

I've spent over a decade doing this inside real organizations. At IBM I designed AI/ML architecture across industries for 9+ years. At Galileo AI I worked with Fortune 500 CTOs on AI reliability and observability. Now at ServiceNow I sit at the intersection of architecture and adoption, figuring out why enterprise AI initiatives stick or die.

Outside of that, I teach enterprise AI through MIT Sloan, UCLA, and UT Austin executive programs, hold multiple AI patents, publish review and judge for IEEE and EAAI, working with 75,000+ professionals across those programs.

Previously at Samsung, IBM, NASA/NOAA

Samsung
NOAA
Direct Supply
UCLA
See all products from Vyoma

Sign up to join this lesson

By continuing, you agree to Maven's Terms and Privacy Policy.