Today, we're excited to introduce Conduit: the Industrial Context Mesh that adds meaning without movement.
The Problem with Industrial Data
Industrial organizations face a fundamental challenge: their operational data is fragmented across dozens of systems. Historians, SCADA platforms, PLCs, MES systems, and enterprise applications all hold pieces of the puzzle. Getting answers requires either:
- Moving all data to a central location - expensive, slow, and creates governance nightmares
- Building point-to-point integrations - creates technical debt and doesn't scale
- Asking data engineers to write custom queries - creates bottlenecks and delays insights
None of these approaches work well. The first two require massive investment. The third means operational insights take weeks instead of minutes.
A New Approach: The Context Mesh
Conduit takes a fundamentally different approach. Instead of moving data, we create a semantic layer that sits above your existing systems and adds context to data where it lives.
How It Works
- Connect - Deploy lightweight adapters to your existing systems (Ignition, OSIsoft PI, OPC-UA servers, Splunk, and more)
- Discover - Conduit automatically discovers tags, points, and metadata, building a unified catalog
- Contextualize - Define semantic relationships using our ISA-95 inspired ontology
- Query - Use natural language or our Conduit Query Language (CQL) to get answers in seconds
No Data Movement
The key innovation is that your data never moves. Conduit's federated query engine executes queries directly against source systems and merges results in real-time. This means:
- No ETL pipelines to build or maintain
- No data lakes to manage
- No governance concerns about data duplication
- No latency from batch synchronization
Natural Language Queries
Perhaps the most exciting feature is our Natural Query Engine (NQE). Instead of writing SQL or navigating complex tag hierarchies, you can simply ask questions:
"Show me the temperature and pressure readings for all fermenters in Building A that exceeded setpoints in the last 24 hours"
NQE understands your semantic model, translates this to optimized queries against multiple source systems, and returns unified results in seconds.
What's Next
We're starting with support for the most common industrial data sources:
- Ignition (MQTT, historian, tags)
- OSIsoft PI (AF, Data Archive)
- OPC-UA servers
- Splunk (log data)
- SQL databases
We're also working on adapters for Aveva, Rockwell FactoryTalk, and Siemens MindSphere.
Get Started
Interested in seeing Conduit in action? Request a demo and we'll show you how to get unified insights from your industrial data in minutes, not months.
Conduit is currently in private beta. We're working with select manufacturing and process industry partners to refine the product before general availability.