The Ontology: Resilience in Crisis

The Palantir Foundry Ontology enables organizations to withstand crisis — from war in Ukraine, an energy crisis in Europe, inflation, to rising tensions in the Pacific.

In the last few years alone, the Ontology has powered:

• Support for refugee relief across Europe amid the war in Ukraine
• Evacuating 15,000 Afghan refugees on 24 hours’ notice
• Elective waiting list management for Chelsea and Westminster Hospital NHS Foundation Trust, helping patients get seen faster
• Power utilization optimization for power-intensive industries like manufacturing and wastewater treatment amid rising energy prices
• Emergency wildfire response for California utilities like PG&E
• A strategic shift toward balancing production and net zero ambitions in a profitable way for bp

In this blog post, we explore the Ontology, and how it’s applied in practice to help make organizations more adaptable.

Why we built it

Let’s start with where the Ontology came from. We saw that enterprise software was falling short for our customers in their most critical moments. When they needed to adapt, their software stood in their way.

• Our utilities customers had grid software that was meant to keep the lights on. When wildfires demanded they turn off the lights, their software proved to be rigid and unadaptable.
• For our manufacturing customers, what happened when a demand-constrained world became supply-constrained? What happened when every piece of software organizations had assumed supply was available on demand and lead times were predictable? Things broke.
• In healthcare, what happened when a global pandemic upended hospitals, placing unprecedented strain on personnel and resources and delaying elective care? Lives were at risk.

They had all done the “right” things — brought data together, created golden tables, created risk models, visualized. They often told us, “I got all my data in one place. But it’s not helping me when I need it most.”

Time after time, we saw how this linear approach was failing them when it mattered most.

These realizations drove us to develop the Ontology — the key differentiator of our Foundry platform and how we’ve been able to help organizations deliver rapid and meaningful impact.

The Ontology transforms digital assets — including data, models, and processes — into a dynamic, actionable representation of the business for all users to leverage in operations:

• Objects digitally represent the real-world entities, relationships and events, that constitute your business.
• Relations represent the connections between real world entities, events and processes.
• 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.

How it works

It starts with data and models, spread throughout the organization. Data could come from ERP systems, CRM systems, IoT, GIS systems, and even homegrown databases. Models could pertain to demand forecasts, asset monitoring, risk factors and processes driven by AI/ML or optimization.

The Ontology continuously synchronizes these data and models into human-centric objects.

A utility would have an Ontology reflecting transformers, substations, customers, field crews — rather than rows and columns and queries. This makes it possible for non-technical users to make meaning of data and action it in an informed way. More on that here.

We call this “hydrating” the Ontology.

Why you need more than a data model

The Ontology is not a pre-defined data model. It is a software system that enables collaboration across the entire organization. Like a language, it can be manipulated and used to construct increasingly complex structures according to pre-defined rules, and captures relationships and actions between objects. It’s a language that organizations can use to speak to all their existing systems.

The Ontology helps put an end to, for example, multiple versions of the same customer proliferating throughout the organization, while simultaneously defining a common set of actions that can be taken by a range of operational teams.

Having a single digital plain is particularly important in times of crisis, as people struggle to gain an understanding of the ground truth, and take actions to influence it. If you think of an organization as a Venn diagram, some of the most impactful work falls in the intersections between teams:

• Utilities customers need to be able to shut off power without jeopardizing critical operations for their customers, such as hospitals. To do this, they need a common understanding of which transformers correspond to which customers and a linkage between Public Safety Power Shutoff (PSPS) and customers, which are managed by separate teams.
• Manufacturing customers need to fully understand their inventory — which warehouse it’s in, if alternative bills of material are available, and how to allocate it. Teams that manage stock need to connect to those in charge of production and to those with an on-the-ground picture of customer priority.
• Hospitals need to understand and manage their operations under strain, when to take transfers and schedule elective surgeries, how to allocate staffing and beds. The data related to staffing and patients is usually held by different teams across the organization.

A single view of integrated operations is essential to fuel effective collaboration across groups — from IT to on-the-ground-operators.

Why it matters

In the Ontology, data, models, and actions are co-equal primitives: data and models are only valuable if they lead to action. Actions correspond to real world decisions, closing the loop with data and models across the enterprise. When Actions are taken, they are immediately written back to the Ontology and core operational systems, permeating the entire organization.

• Take, for example, a fire event. When crisis strikes, a utility can use the Ontology to see all grid assets and the connections between them — and therefore de-energize the grid at the right place, and at the right time, all while keeping the customer informed.
• Data from PSPS events can inform asset-wide risk models that power inspection prioritization and maintenance schedules into the future. These models can even inform proactive communication to customers if they are living in high-risk areas.

Over time, these actions can be collectively reviewed and used to power learning loops.

By harmonizing data, models, applications, workflows, simulations, and analytics and connecting all of these elements to actions, the Ontology provides a stable foundation on which to build — and learn from over time — no matter how complex the world becomes.

In this way, the Ontology empowers organizations to continuously, seamlessly adapt — no matter the challenge.

Author

Emily Nguyen, Head of Industrials, Palantir

---

The Ontology: Resilience in Crisis was originally published in Palantir Blog on Medium, where people are continuing the conversation by highlighting and responding to this story.