When AI Agents Become Your Users: What B2B SaaS Gets Wrong About the Next 3 Years

Watch the full podcast here: YouTube

This episode tackles one of the most consequential — and underappreciated — shifts in enterprise software: the transition from human users to AI agents as the primary consumers of B2B SaaS products. The panel brings together five voices across product, engineering, finance, and early-career practice to stress-test a single thesis: most B2B SaaS companies are optimizing for the wrong user, and the window to course-correct is narrowing fast.

Gyanesh Samanta and Kamaljit Deka anchor the product side of the conversation — examining what it actually means to architect software for agentic consumption, where the PM's mandate shifts from user empathy to machine-readable design, and how ProductOS's 13-agent orchestration model exposes both the promise and the quality-gate risks of AI-native product workflows. The discussion goes beyond theory, using IBM Sterling OMS and ProductOS's own 0-to-1 lifecycle as live case studies.

Tamonash Pan and Eshaan Bhardwaj bring the financial services lens — an industry where the cost of a wrong AI decision is regulatory, not just reputational. Their conversation surfaces the governance gap at the center of this shift: auditability, liability, and intent-tracing aren't compliance checkboxes, they're the product features that determine whether an AI-native tool is enterprise-ready or just a well-funded demo.

Ritika Dash closes the loop from the early-career PM perspective — asking the harder question underneath all of it: when AI tools automate discovery, PRDs, and execution, what does genuine product instinct look like, and are we building the next generation of PMs or just the next generation of AI prompt managers? Together, the panel constructs a sharp, opinionated map of where B2B SaaS is headed — and what builders, buyers, and practitioners need to do differently in the next 36 months.