The Ontology: Your Business as Code

What Is the Ontology?

The Foundry Ontology is a "digital twin" of the entities, events, and processes that are important to an organization . It transforms your digital assets—including data, models, and processes—into a dynamic, actionable representation of the business that all users can leverage in operations .

In simpler terms: it's your business, codified.

The ontology creates a complete picture of an organization's world by mapping datasets and models to object types, properties, link types, and action types . This approach enables a more intuitive and meaningful way to manage and utilize data .

The Three Pillars of the Ontology

The ontology is built on three foundational concepts that mirror how the real world works:

Objects

Objects digitally represent the real-world entities, relationships, and events that constitute your business. An object can be a customer, a policy, a claim, a vendor, a purchase order, a piece of equipment—anything that matters to your operations .

Relations

Relations represent the connections between real-world entities, events, and processes. For example, a Purchase Order object might link to a Vendor object and a Material object, showing how these entities interact . These codified links allow business users to explore data and navigate through entities without any technical knowledge or complex programming languages .

Actions

Actions capture the kinetics between objects and orchestrate real-world change through your enterprise systems. Actions can be mapped from existing processes or existing models. In practice, this means you can click a button to "Approve Purchase Order" or "Flag Claim for Review," and the action writes back to your source systems .

This is what makes the ontology two-way: IT can feed it with data from transactional systems and also configure writeback actions that update the source systems. End-user applications are fed from the ontology in a user-friendly way and, when needed, trigger these writeback actions .

Why the Ontology Matters: Breaking Down the Business-IT Wall

For most of my career—prior to joining Palantir—I held various CTO roles where I was ultimately responsible for sourcing and implementing the right technology so organizations could achieve their strategic goals .

I often reflected during those years that an organization's IT architecture system often mirrors the organization's structure. Looking around your own organization, how many departments co-own the same IT system? It can be incredibly challenging to source a system that is consistently owned by both IT and the business .

Most enterprises define the tools they use into one of two general buckets: enterprise systems or end-user applications. The former are typically owned by IT and the latter mostly used by business end-users .

The main rationale behind this division is that organizations have clear owners and SLAs on each side. However, its downfall manifests with larger IT projects which don't have direct business value, leaving business stakeholders often feeling like they don't have the tools and data to make informed decisions .

Enter Palantir Foundry.

Foundry breaks down that wall. Co-owned by both IT and business users, Foundry helps IT teams with their data integration efforts while providing business units the tools and applications they require to confidently iterate with their data . This positive disruption is made possible through ontological modeling .

How Ontology Hydration Works

"Ontology Hydration" is the process of combining data, models, and transactional systems with intuitive tooling built to empower both technical and non-technical users.

Pipeline Builder

Foundry's flagship Ontology Hydration capability allows you to build production-grade pipelines using a dynamic point-and-click interface. You start by defining endpoint schema for Ontology object types and properties and describing the pipeline to match inputs to endpoints. Pipeline Builder's backend then automatically writes transform code to build the desired pipelines.

Pipeline Builder works with all types of data—structured, unstructured, IoT, geospatial—across all modalities (batch, micro-batch, streaming) .

Ontology Manager

Ontology Manager allows you to design your Ontology with maximum precision to fit the context of your organization. You can easily define your organization's Objects, as well as their Links, Properties, and metadata .

You can specify the key decisions and behavior in your organization, with rails for mapping models to Actions that users or systems can implement—closing the loop between AI/ML and operations .

The Future: Your Business as Code

The concept of an AI operating system—a full software and hardware stack that brings production-grade AI agents and workflows to organizations—is rapidly becoming reality. This has been demonstrated through recent partnerships, such as Palantir and Dell's joint solution with Foundry and the Ontology platform running on-premises inside Dell AI Factory with NVIDIA .

The result is what they're calling an AI operating system—a full software and hardware stack that brings production-grade AI agents and workflows to organizations that can't afford to let their data leave the building .

This isn't a partnership of convenience. It's a direct answer to a problem that has quietly stalled AI adoption at the highest levels of enterprise and government: you can have powerful AI, or you can have control. Until now, you couldn't easily have both .

As Scott Spradley, CTO of Tyson Foods, put it: "When you start thinking about the power of ontological modeling, our mindset was when you start to couple that together with the semantic layer that Palantir has, you've got a huge win opportunity. Massive. When you can start to make sense out of vast amounts of data that seemingly are just totally disconnected, and you can find relevance between those, you're in a very different spot" .

Real-World Impact

The ontology is already transforming critical industries:

Insurance

Mexican insurer GNP has become Palantir's first publicly announced commercial customer in Mexico, using AIP to unify data across claims, underwriting, and operations into a single, coherent Ontology. Building on that foundation, AIP enables GNP's teams to deploy AI agents and models that connect fragmented data, surface risk information that previously required hours of manual work, identify anomalous claims patterns, and flag potential fraud before payments are disbursed. The partnership has extended into underwriting, where GNP is building a bespoke intelligent decision engine to replace legacy off-the-shelf systems in which a single rule change could take months to approve and implement .

Mining

Rio Tinto, an early adopter of Foundry, primed its operational landscape for AI deployment through the creation of a robust digital twin. Via the Ontology, AIP enables Rio Tinto to build, test, and validate AI use cases at an accelerated pace and deploy them to production safely .

For Network Specialists and Train Controllers in the RTIO Operations Centre in Western Australia, Foundry provides a comprehensive view of Rail operations, assembled from real-time data from hundreds of equipment units and systems in the value chain. With the Ontology providing a unified view of all assets, Network Specialists coordinate the haulage of iron ore by 53 driverless trains, each with 240 wagons, across the Pilbara rail network .

In Mongolia, Foundry equips Rio Tinto with a dynamic understanding of geotechnical risk at Oyu Tolgoi, one of the world's deepest and largest block cave mines. The Ontology integrates data from thousands of sensors across the mine and serves as a single source of information for cave health, instrumentation, and risk. This represents a new paradigm for block cave mining .

Specialty Insurance Underwriting

AIG and McGill and Partners entered into a collaboration using AI to manage the deployment of insurance capacity across the broker's specialty portfolio. AIG worked with Palantir to build an ontology of the broker's portfolio using Foundry, translating raw data into a structured map of how entities, risks, and relationships connect, allowing AI systems to reason about them .

The Bigger Picture: Why This Matters for You

When you start thinking about the power of ontological modeling, you begin to see that the future of enterprise operations is not about having more data—it's about having data that understands itself.

The ontology is the foundation on which end-user applications are built by the business, while connecting back to any IT-owned transactional system for both reads and writes . This "two-way street" model—where IT feeds the ontology with transactional data and configures writeback actions, while business users build applications that trigger those actions—is redefining the relationship between technology and operations .

The question isn't whether you should build an ontology for your business. It's how quickly you can do it—before the competition leaves you behind.