SAP CEO: AI Won't Kill Software, But It Will Change Your Job — With Christian Klein
Summary
SAP CEO Christian Klein discusses the state of enterprise software, arguing that AI agents require 100% accuracy and deep business context to replace legacy systems. He explains that while LLMs are powerful, they must be integrated with SAP’s data ontology and governance frameworks to handle critical tasks like financial closing. The conversation covers workforce transformation, the limits of "vibe coding" for complex ERP systems, and why SAP’s platform remains essential despite the rise of open-source models and third-party AI interfaces.
Topics discussed
Intro: Software boom post-saspoalypse and sponsors
The 'Mountain' analogy: AI accuracy limits and business value
Overcoming obstacles: Structured data, governance, and 100% accuracy
Data quality, silos, and the role of AI in cleaning enterprise data
Accelerating release cycles and cultural shifts in software companies
Workforce impact: Burnout, job security, and the nature of work
Reskilling programs and the shift from rote to inventive tasks
The 'Saspoalypse' myth: Why vibe-coding won't replace enterprise software
SAP's moat: Domain knowledge, ontology, and agent governance
Sponsor breaks: Notion, Avpoint, and Ironwall
Platform strategy: Building agents on SAP vs. third-party models
Model selection, cost optimization, and the agent gateway
Market response: SAP stock recovery and shifting AI spend trends
The rise of table AI models and specialized prediction tools
Managing token spend: Limits, model switching, and TCO
Multi-agent swarms and security vulnerabilities in AI development
European regulation: AI Act, GDPR, and the 'Omnibus' simplification
Closing remarks and sign-off
Listen ad-free on Castria