Insights and Key Learnings on Our AI Journey
summary
Company converts five years of pilots into an enterprise data platform with Databricks feeding Google Cloud Storage BigQuery Tableau CA, builds a reusable AI-as-a-service template on Vertex AI and Zenal plus Docker, runs 4–5 PoCs for document OCR and a license‑classification model that auto‑routes licenses when confidence is high and falls back to human review, institutes 4‑hour discovery workshops, adds automated data‑product quality checks, and still wrestles with SAP integration, data quality limits and nondeterministic LLM behavior.
AI-written summary · about
chapters (15)
- 0:00GEMA’s AI journey intro
- 1:09Meet the team and GEMA’s role
- 4:00GEMA like a jam band with AI
- 6:09Early AI challenges learned
- 8:18Mindset and digital transformation
- 11:51Training teams on AI fundamentals
- 14:19Growing AI expertise and updates
- 14:39Enterprise data platform for AI
- 18:35AI infrastructure as a service
- 21:04Barriers in AI project adoption
- 22:31AI use case discovery workshop
- 27:33Success and unified testing concepts
- 29:39AI for document processing
- 33:12Key AI improvements and initiatives
- 34:36Q&A on evaluation and data quality
related talks
17mHow will AI transform your data platform?
56mAI War Stories: Real world learnings delivering enterprise AI solutions
58mAI in Reality Fireside Chat: Enterprise AI & Open‑Source Innovation
01mDatabricks Customers
37mOn Building Production-Quality AI Systems
11mUsing AI to extend the range of data scientist
19mDatabricks
27mAI as a Service Scaling AI with Ease and Innovation