Enabling Industrial AI with Validated Data: HighByte on
summary
Bayer connects HighByte to AVEVA PI and AWS Bedrock to let an LLM query validated pharma data without copying whole historians, exposing only event frames and aggregated asset framework views via Model Context Protocol so the model pulls precise batch yields and comparisons; CSV and legacy file quirks get cleaned in-flight; a controls engineer runs a chatbot with human-in-the-loop validation, proving remote, repeatable answers across heterogeneous factories and cutting dozens of static reports and manual clean‑room lookups.
AI-written summary · about
chapters (6)
- 0:00Introduction and AI challenges
- 3:08Why industrial AI is hard
- 5:57HighByte and industrial data ops
- 8:23Bayer’s AI use case with PI system
- 11:34Demo and future of AI in industry
- 12:37Scaling AI across factories
related talks
45mOur way to a data-driven enterprise - A Journey by Bosch
21mShaping the future of health: Boehringer Ingelheim's Data & AI Journey
31mAI, Digital and Value Streams Enable Data Driven Organizations
11mDemystifying AI in Smart Manufacturing
15mAI Gardening: Harvesting and Pruning the AI Tree
27mBuilding Reliable Human-AI Warehouse Operations at Scale
27mAI at Scale in the Enterprise
23mCognitive Architecture for Process industries