Interview

Why your AI pilot never ships

Christian Monberg

CTO & Head of Product, Zeta Global

69 minHosted by Kim Isenberg and Peter Thum

Watch on YouTube

About this conversation

Why do so many enterprise AI pilots work in a demo but never become part of everyday business? Christian Monberg, CTO and Head of Product at Zeta Global, explains to Kim Isenberg and Peter Thum what gets in the way: fragmented data, isolated tools, and workflows that nobody has redesigned around AI.

They discuss Zeta’s work with Palantir Foundry, how its Athena system connects recommendations to data and forecasts, and why agents still need human checkpoints before actions such as large-scale data merges — plus how to measure trust, choose between open-source and frontier models, and weigh the cost of moving onto a new platform.

Disclosure: Zeta Global covers the video editing costs.

Topics

  • Why AI pilots stall
  • When systems disagree about the customer
  • Zeta’s integration with Palantir Foundry
  • Where AI agents need human approval
  • How to build trust in AI and test agents before they act
  • Open-source versus frontier models
  • Where to start in the next 90 days

Chapters

  1. 0:00Highlights
  2. 1:11Introducing Christian Monberg
  3. 3:13Why AI pilots stall
  4. 4:17Lessons from building Boomtrain
  5. 6:39From marketing cloud to AI infrastructure
  6. 8:33What Zeta helps customers do
  7. 11:09When systems disagree about the customer
  8. 13:08Zeta's integration with Palantir Foundry
  9. 15:49Data problems, model problems and adoption
  10. 19:12Changing how marketing teams work
  11. 24:09From incoming data to business action
  12. 25:56How Athena grounds its recommendations
  13. 28:14Ontologies and the meaning of business data
  14. 33:01Why integrate with Foundry?
  15. 34:25Where AI agents need human approval
  16. 37:54How do you build trust in AI?
  17. 41:08Testing agents before they act
  18. 43:55Measuring results and connecting workflows
  19. 47:15Is AI saving time on the right things?
  20. 50:25Open-source models versus frontier models
  21. 52:33Migration costs and time to value
  22. 55:58Adapting to each customer's systems
  23. 57:07Switching vendors and avoiding lock-in
  24. 59:11Where to start in the next 90 days
  25. 1:01:22Competing as AI becomes mainstream
  26. 1:05:07Motorcycles, architecture and lasting design
  27. 1:07:43Where to follow Chris and Zeta

Kim Isenberg is co-founder and Editor-in-Chief of Superintelligence, an AI newsletter published every weekday.

Read the Newsletter

More interviews

All interviews