The Reasoning Show
The Future of Service belongs to Self-Improving AI
Apr 01, 2026
Massive Studios
SUMMARY: Today’s episode is all about a transformation happening in customer service—one that’s moving us from static systems and scripted workflows into something far more dynamic: AI systems that can actually learn and improve over time.
GUEST: Shashi Upadhyay (President of Product, Engineering, and AI at Zendesk)
SHOW: 1015
SHOW TRANSCRIPT: The Reasoning Show #1015 Transcript
SHOW VIDEO: https://youtu.be/IQaxE-DjIpo
SHOW SPONSORS:
SHOW NOTES:
Topic 1 - Welcome to the show. Tell us a bit about your background and your focus today.
Topic 2 - You describe this moment as a shift from systems of record to intelligent systems of action. What’s fundamentally broken in today’s customer service model that’s forcing this transition now? What changed in the last 2–3 years to make this possible?
Topic 3 - There’s been a lot of AI in customer service that overpromised and underdelivered. What are the biggest gaps between what customers actually need—like resolution—and what legacy automation has been delivering?
Topic 4 - The concept of a “self-improving” system is really powerful. What’s actually new here—what enables AI to improve with every interaction without constant human tuning?
Topic 5 - You’ve moved from assistive copilots to what you call “agentic AI” that can resolve issues end-to-end. Where are we today on that journey—and what still requires human involvement?
Topic 6 - Voice has historically been one of the hardest channels to automate. What changes with this new generation of AI that makes even complex, multi-step voice interactions solvable?
Topic 7 - If we fast-forward 2–3 years, what does a “best-in-class” customer service experience look like in an AI-first world?
FEEDBACK?