Keep pulling the thread on Philipp Herzig.
SAP's "Joule for Consulting" product can reduce the effort required by consultants on complex projects by 30%, leading to faster outcomes and lower costs.
SAP developed a model named RPT1 (Relational Pre-trained Transformer) specifically for predictive tasks on structured, tabular data, and has published research on it at NeurIPS.
AI is driving a business model transition for SAP away from seat-based licensing and towards consumption-based pricing.
SAP is the market leader in enterprise software applications and platforms, serving 400,000 enterprise customers.
Philipp Herzig observes a growing gap in enterprise AI between the "innovation race" and the "outcome race," indicating that business adoption is not keeping pace with technological progress.
All developers at SAP are using agentic coding to boost productivity across all programming languages.
SAP's "Joule for Consulting" is one of the company's fastest-growing AI products.
SAP has launched AI-powered travel booking and expense agents within its Concur product line.
Philipp Herzig believes the era of designing static user interfaces that require a human to learn them is over, and SAP is moving towards dynamically generated UIs.
The biggest engineering challenge for SAP in AI is not building the models, but teaching them to perform correctly and reliably at enterprise scale.
SAP's platform has 20,000 APIs, which presents a significant scaling challenge for AI agents that rely on tool use due to issues like context window limitations.
Philipp Herzig believes that Test-Driven Development (TDD) is making a comeback because AI agents require clear, testable outcome descriptions (evals) to be developed effectively.