SAP CTO Philipp Herzig thinks AI coding tools miss running business
SAP’s chief technology officer Philipp Herzig argues that AI coding tools, while able to spin up apps quickly, overlook the harder task of running those apps in a business context. He says the “SaaS is dead” narrative conflates app creation with production‑ready operation, and that many pilots fail because they lack the governance, security and business‑process integration needed at scale.
Key points
- Joule Work lets users state intent, assembling data from multiple apps into a visual workspace
- 60,000 SAP colleagues use Joule Work internally, and SAP has committed over €1B to Prior Labs
- Tabular AI model SAP‑RPT‑1.6 offers explain‑by‑feature and explain‑by‑example to build trust in predictions
Herzig highlights SAP’s new Joule Work platform, which lets users state intent instead of clicking screens, and Joule Studio, a tool that gives developers a pre‑built enterprise context for building agents. He also discusses SAP’s Knowledge Graph and Company Memory, which map business data and capture decision rationale to reduce “fluent‑but‑wrong” answers. The company has committed more than €1B to Prior Labs and built SAP‑RPT models, with 60,000 colleagues using Joule Work internally.
Key numbers include 80% of AI users feeling more productive, only 6% seeing profit impact, and a 50‑to‑45‑day reduction in average collection time from a tabular AI use case. Herzig stresses that the hardest engineering challenge is building safe, scalable agents with proper guardrails, not just making them work.
SAP's Philipp Herzig on AI built for business data
The Rundown AI · 25 October 2026
Loading the full article…
This text was published by The Rundown AI and written by Melina Panitsidis. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Enterprise & Industry
All →- Boeing and Microsoft build AI-powered Dataliner platform for airlines · 1 src
- Claude Opus 5.5 may change default effort and API settings, reports say · 1 src
- Peach Tech starts company-wide Claude certification program for all employees · 1 src
- AWS says most AI cloud capacity booked through 2027, many 2028 slots filled · 1 src
- Google Vault adds filtered exports for Gemini in Workspace · 2 src
Comments
via GitHub Discussions