Industrial AI Podcast Bridging the gap between factory data and agents
13 snips
Feb 4, 2026 John Harrington, Co-founder and CPO at HighByte, a leader in Industrial DataOps. He breaks down MCP and why curated, contextualized data matters for connecting factory systems to intelligent agents. They talk MCP vs OPC/MQTT, HighByte’s agent features, tool curation to reduce hallucinations, and how data orchestration scales access from historians to cloud services.
AI Snips
Chapters
Transcript
Episode notes
HighByte's Data-Ops Origin
- John Harrington described HighByte as an industrial data-ops company that curates and moves OT data into IT and cloud systems.
- He emphasized contextualizing data for target consumers rather than moving raw bytes.
MCP Bridges LLMs And Industrial Data
- MCP (Model Context Protocol) was designed to let LLMs request and receive contextualized industrial data via described 'tools'.
- It maps human-friendly descriptions and parameters to data access, making LLMs effective consumers of OT/IT data.
Use AI To Build And Explain Pipelines
- Provide both AI-for-dataops (assist users) and dataops-for-AI (deliver curated data) features in your platform.
- Expose pipelines and let users create and inspect them via LLM-driven interfaces to accelerate onboarding.