ProductCon NY 2026: How winning SaaS companies are transitioning to AI-native

This panel from ProductCon New York 2026 (May 20) brings together product leaders from four enterprise SaaS companies—Birkan Icacan from Enterpret, Aditya Ganjam from Conviva, Vijay Umapathy from Contentsquare, and Phillip Badger from Lucid Software—for a conversation about what transitioning to AI-native actually means inside an existing B2B product. Rebecca Geraghty of Publicis Media moderates.

The framing rejects the idea that AI will structurally eliminate SaaS subscription businesses. The panelists argue that businesses continue to value predictable pricing, governed infrastructure, and trusted vendors—and that enterprise requirements around compliance, security, and uptime are becoming more important as AI deployments scale, not less. The conversation focuses on the messy middle: what changes in product strategy, team structure, and delivery when you are not building an AI-first startup from scratch but adapting an existing product with an existing customer base.

The panel is most useful for PMs and product leaders at established SaaS companies who need frameworks for explaining the transition pace to stakeholders, or for PMs working on enterprise AI features where governance and context are first-class requirements.

Key takeaways:

1. The competitive moat has shifted from features to governed infrastructure. Badger made the case that when source code is cheap or free, the differentiator is whether a company can deliver a trusted, governed, and context-aware environment for high-stakes decisions. Enterprises evaluating AI vendors are increasingly asking about data provenance and access control rather than demo performance.

2. Products now have two distinct types of users. Icacan shared that at Enterpret, one-third of users now engage via an MCP server rather than the UI. This means product teams need to design for machine access as a first-class use case from the start, not retrofit it afterward. APIs and CLIs have become primary interfaces for a growing segment of real users, not just developer add-ons.

3. AI has changed who can build, but not what enterprise deployments require. The panel agreed that prototyping is now accessible to a much broader audience through vibe coding and low-code AI tools, which has changed how teams explore solutions. However, building enterprise-grade systems—with governance, performance at scale, and integration complexity—remains a different category of work from demo to production.

4. The roadmap question has changed from prioritization to layer identification. Rather than prioritizing features against each other, the strategic move the panelists recommend is identifying which product layers are being commoditized by AI and which remain proprietary. Teams that are honest about this distinction can concentrate investment on what will still be defensible in three years.

Worth watching if you are responsible for an AI feature roadmap at a company with existing enterprise customers and need concrete frameworks for explaining to leadership why the AI-native transition takes longer and costs more than the prototype suggested.