"The AI Race Won't Be Decided By Models & Tokens, But By Economics"

PHILIPP HERZIG, Chief Technology Officer at SAP, discusses the changing mandates of tech leadership, the next phase of AI adoption, SAP's rapid progress with Joule and RPT1, and why India is core to SAP's future, in an interaction with BW Businessworld's Rohit Chintapali. Excerpts

- Rohit Chintapali

Your role was formally defined as Chief AI Officer in 2024, and then you took on the CTO role in 2025. Today many companies are appointing CAIOs. How is the CAIO role distinct, and how does it differ from the CTO mandate?

It depends on how the role is defined, but broadly, a Chief AI Officer needs to look at AI end-to-end, not just from a technology standpoint, but also business, go-to-market, marketing, services and internal adoption. It’s a truly 360-degree view, because AI touches everything.

A CAIO does not have to be the world’s best AI researcher. They need a strong understanding of AI, of course, to hire the right people and make the right decisions. But I’ve seen brilliant AI researchers struggle with business elements. On the other hand, great general managers may lack the technical depth required for AI. Finding someone who sits at that intersection is extremely hard.

The CTO role goes beyond AI. Yes, AI is a major part of what I do, but the CTO must also look at data, integration, identity management, process orchestration and the boundary conditions required for AI to be successful. We must think about SAP’s “North Star” architecture for the next three to four years and steer investments across platform, data and emerging technologies — including quantum, robotics, sovereignty and trusted cloud.

As someone who has led SAP’s AI agenda for nearly two years now, what feels different about 2025? And where do you see the momentum heading into 2026?

2025 has been the year of adoption. We built the foundations earlier, and this year we had to build on top of them, while also disrupting ourselves with the shift toward agents.