- Moving from 'Human-in-the-loop' to 'Agent-as-the-loop' by using decision traces to continuously update enterprise standard operating procedures.
- The necessity of specialized 'tabular Transformers' (RPT) because standard LLMs struggle with the regressive and classification requirements of enterprise financial planning.
- Why AI adoption in the enterprise is failing to scale: the lack of clean, harmonized data models and the difficulty of securing agent access to massive, complex API ecosystems.
Channel: No Priors: AI, Machine Learning, Tech, & Startups
Source Video
SAP CTO Philip Herzig on Enterprise AI Strategy and Industrial Scale
SAP CTO Philip Herzig discusses the company's shift toward agentic AI workflows, the challenges of implementing large-scale enterprise automation, and the limitations of LLMs in predictive and tabular data tasks.
Key Takeaways
- Conventional UI-driven software is moving toward generative, proactive systems that operate autonomously to execute end-to-end business outcomes.
- Enterprise AI adoption is hindered less by model capability and more by the challenge of managing fragmented data landscapes and ensuring rigorous, verifiable output quality at scale.
- The industry is shifting from 'chatting with data' to deep integration where agents act as autonomous stewards of complex, structured business processes.
Talking Points
Analysis
This conversation is critical for enterprise architects and AI strategists navigating the transition from 'AI hype' to 'AI outcome...
Full analysis available on Pro.
Time saved:
Channel: No Priors: AI, Machine Learning, Tech, & Startups