Empowering Business Users to Apply AI to SAP Data
summary
Shows concrete pipelines: SAP tables are turned into data products (single flat tables) fed to Fabric and Power Platform, clustering cuts 3,500 invoices-without-PO into eight actionable groups, and a hybrid pipeline (clustering + LLM) generates readable explanations via an AI analyst. Demonstrations include process mining on 1.9 million variants and a 150,000-case Italian fines dataset collapsed to 231 variants, plus notes on retraining OpenAI models, Azure connections, and concrete failure modes like watermark-driven misclassification.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 3:56AI adoption challenges
- 8:11Black-box AI risks
- 10:20SAP data integration
- 13:15Semantic layers for users
- 16:02Invoice rule breach detection
- 20:19AI analyst insights
- 23:59Process mining workflows
- 29:47AI process analysis demo
- 35:06Key learnings
- 37:58Q&A
related talks
07mLeveraging AI & Predictive Analytics in Enterprise Architecture
15mBeyond the Silo: Modernizing SAP with Snowflake on AWS for AI
27mEmpowering Innovation through DevOps & AI Integration
17mHow will AI transform your data platform?
22mThe AI Adoption Gap – How We Accelerate AI Use in Enterprises
44mArtificial Intelligence Reloaded - AI Apps in the Industry
45mHarnessing Data Science and AI in Fabric
56mAI War Stories: Real world learnings delivering enterprise AI solutions