AI-Ready Data on Databricks: How TK Elevator Uses Context and Meaning to Make AI Agents Work

Description

How do you bring AI to a decade of fragmented data across one of the world's largest elevator companies? In this episode of The Data + AI Exchange, Viktoria Semaan sits down with Matthias Gohl, Chief Digital Officer, and Christian Jung, VP of Digital, Data & AI Foundation at TK Elevator, to break down how they turned siloed IoT telemetry into agentic AI, processing 500 million data events a day to predict failures before they happen. They get into the real work behind AI-ready data: consolidating a fragmented data lake with Unity Catalog and Medallion architecture, building a trusted and auditable foundation, and giving data a common meaning so both engineers and AI agents can use it. Plus conversational analytics with Databricks Genie, applying LLMs and RAG to unstructured contracts and service logs, and how intelligent agents now support field service and portfolio decisions with humans in the loop. ⏰ TIMESTAMPS 00:00 – Bringing AI to Physical Infrastructure 01:21 – From IoT Telemetry to Agentic AI 04:14 – Predictive Maintenance at 500M Events a Day 06:28 – Data-Driven Field Service Workflows 08:00 – Harmonizing Siloed Data with Unity Catalog 09:49 – Building a Trusted, Auditable Lakehouse Foundation 11:35 – Conversational Analytics with Databricks Genie 12:23 – RAG and Agentic AI for Unstructured Data 14:33 – The Roadmap for Agentic Orchestration 15:11 – Advice: Engineering AI-Ready Data 🔗 RESOURCES Go deeper: TK Elevator's Data + AI Summit session on turning fragmented data into data-driven service: https://www.databricks.com/dataaisummit/session/how-tk-elevator-turning-fragmented-data-digital-data-driven-service Try Databricks Free Edition: https://bit.ly/dbx-free-signup #Databricks #AgenticAI #UnityCatalog #PredictiveMaintenance #TKElevator