IRIS Foundry · Cortex

AI that understands your enterprise, not just your data

Enterprise data sits in separate systems, so people and AI never see how anything connects. Cortex Graph brings assets, customer orders, documents, labor records, and shifts into one connected view: a knowledge graph generated from your Unified Data Model (UDM). Cortex Agent reasons across those connections to answer questions that span systems, instead of looking data up one system at a time.

Predictive maintenance, agentic workflows, connected-worker copilots: every initiative stalls when the AI behind it cannot tell how an asset, an order, and a shift relate. Cortex Graph gives those systems a structured, governed model of your operation to reason against, and Cortex Agent puts it to work in plain language.

Cortex Agent takes a question in plain language, finds the related assets, orders, documents, and shifts across your Unified Data Model, and renders the connections it used to reach an answer. What took a chain of lookups across separate systems happens in one conversation.

Every object on screen traces back to its source through a governed pipeline, so the answer arrives with its evidence attached. That is the difference between a demo and something your operations team relies on.

Cortex moves past search into the work itself: one connected view of the enterprise, answers you can trace, an agent that understands context, and analysis it runs itself.

A digital representation of the enterprise

Cortex Graph brings assets, customer orders, documents, labor records, and shifts together in one connected view, generated from your Unified Data Model. Filter to the entity types that matter, open any object to see its properties, and follow the relationships instead of guessing at them.

Animated connected graph with a plant at the center, linked to suppliers, quality systems, safety procedures, and maintenance records, with a panel for filtering by entity type
Trusted data behind every object

Every object connects back to its underlying data through governed pipelines. When Cortex answers a question, your team sees where the information came from. That traceability is what separates an insight people act on from one they have to verify first.

AI that understands context, not just data

Ask Cortex Agent a question in Copilot and it navigates the connections to find related assets, channels, and records, then combines them into one answer. The graph renders the nodes and relationships behind that answer, so you can keep pulling the thread.

AI that goes beyond what is already known

When a question needs analysis rather than a lookup, Cortex Agent retrieves the data from the graph, writes and runs the code in a secure environment, and brings the result back into the conversation. One thread from question to answer.

Cortex Graph is built to be explored visually and queried in plain language, and the two stay in sync.

Interactive graph exploration

Every entity type carries its own color and standardized edge styling, so relationships read at a glance. Adaptive zoom keeps entity names legible as you move through larger graphs.

Filter by UDM type

Display only the nodes relevant to the question in front of you. Filter the graph to specific Unified Data Model types, search types from the navigation pane, and return to recently viewed types in one click.

Focus mode

Double-click any node to reveal its connections and neighboring nodes. The graph centers on your selection as you explore, and returns to the previous view when you exit.

Natural-language search

Full-text search across assets and time-series channels resolves what a question is actually about, matching names, titles, descriptions, and alternate identifiers, then ranking results by relevance.

Copilot and graph in sync

Ask Cortex Agent a question in Copilot and the graph renders the relevant nodes and relationships automatically. Open the graph from a response and it keeps the query context, so follow-up questions stay consistent across both.

API access

Node types, node listings, neighboring nodes, and projected graph queries are all available through the Cortex Graph API, with attribute-level projections that keep payloads small for the systems you build on top.

12 weeks
Time to first value on IRIS Foundry. Cortex Graph is generated from the Unified Data Model you stand up during that program, not a separate modeling project.
300+
Pre-built agents available across industrial operations on IRIS Foundry. Cortex Agent gives them a connected view of the enterprise to reason against.

Straight answers for data, engineering, and operations teams evaluating Cortex as part of an industrial AI program.

What is Cortex, and how is it different from enterprise search?
Enterprise search returns documents that match a query. Cortex holds a connected model of your operation, so Cortex Agent follows relationships between assets, orders, documents, labor records, and shifts to assemble an answer from several sources. When the answer needs calculation, the agent runs it rather than handing back a link.
How is the graph built? Do we model it ourselves?
Cortex Graph is generated directly from your Unified Data Model. The entities, relationships, data, and properties defined in the model become the graph, so there is no separate graph modeling project to fund or maintain.
Can we trace an answer back to source data?
Yes. Every object in the graph connects back to its underlying data through governed data pipelines. That lineage is what lets teams verify a Cortex answer instead of accepting it.
How do people use it day to day?
Two ways, and they stay in sync. Ask Cortex Agent a question in Copilot and the graph renders the relevant nodes and relationships automatically. Or start in Cortex Graph, filter to the Unified Data Model types you care about, and double-click into focus mode to follow connections node by node.
Can Cortex answer questions that require analysis, not just retrieval?
Yes. When a question needs calculation or comparison, Cortex Agent retrieves the relevant data from the graph, writes and executes code in a secure environment, and returns the result in the conversation.
Is there API access for the systems we build?
Yes. The Cortex Graph API covers node types and counts, node listings with pagination and attribute filters, neighboring nodes for incremental expansion, and projected graph queries with field-level selection.

One unified data foundation, one way to build new AI applications, and one layer of agentic execution, built to work together across your plant.

IRIS Foundry

The industrial DataOps platform that unifies DCS, LIMS, historians, MES, and ERP data into one governed, AI-ready foundation.

Explore IRIS Foundry >
IRIS Forge

Turn operational requirements into production-ready AI applications from a natural-language prompt, no custom code, no long development cycle.

Explore IRIS Forge >
IRIS Flows

Put pre-built AI agents to work inside your process, reasoning from context, coordinating with operators, and executing under governed autonomy.

Explore IRIS Flows >

Whether you are standing up a Unified Data Model or already running agents on IRIS Foundry, Cortex gives them the connected view they reason on. Talk to someone who knows industrial data, not a generalist.