In this guest blog post, Tanuja Gupta, Manager: Knowledge Graphs and Explainable AI at MAN and previously Solutions Architect and Knowledge Graph Ambassador at Scania (both part of TRATON GROUP), explains how knowledge graphs have helped Scania—a world-leading provider of transport solutions— and TRATON GROUP—one of the world’s largest commercial vehicle manufacturers—overcome data challenges and create a more connected, consumable, and actionable data environment across the enterprise and its sister brands.

 

Leveraging Knowledge Graphs at Scania and TRATON

In today's data-intensive landscape, efficiently managing and leveraging the immense volumes of data generated across diverse business systems poses a significant challenge. At Scania, we are tackling this issue by implementing Knowledge Graphs (KGs). 

 

Scania is a world-leading provider of transport solutions, including trucks and buses for heavy transport applications, and belongs to the TRATON GROUP, a subsidiary of the Volkswagen Group and one of the world's largest commercial vehicle manufacturers. 

 

These graphs are intended to integrate data not only from Scania but also from other TRATON group companies like MAN, Navistar, and Volkswagen, formatting it in a way that is both machine-readable and understandable. This advancement empowers our systems to interpret and analyze information autonomously, mimicking human-like decision-making processes. By doing so, it significantly enhances our capacity for informed decision-making and boosts operational efficiency throughout the organization.

 

Table of contents

 

• What is a knowledge graph?

• Challenges in data and knowledge management

• Supporting Scania’s information architecture

• Use cases for knowledge graphs at Scania

• Key considerations for building the knowledge graph

 

What is a knowledge graph?

From Scania’s perspective, a knowledge graph is a structured representation of our internal knowledge. Because a knowledge graph is driven by an underlying ontology that captures meaningful descriptions about the concepts used in the data and the relations between these concepts, the knowledge representation is contextualized and context becomes machine-readable and, crucially, machine-understandable. The knowledge graph allows various systems to not only access and interpret the information but also deduce new facts and make decisions based on it—just like a human would. It integrates diverse data sources into a central hub, facilitating automated processes and enabling consistent, accurate decision-making across different systems and use cases.

 

Challenges in data and knowledge management

At present, much of Scania’s knowledge sits across various isolated systems, making it hard to manage and keep synchronized. Repetitive copying of data from one system to another to deduce facts is cumbersome, leading to inefficiencies and errors. Maintaining multiple point-to-point integrations adds further complexity. The ideal state is to have a central knowledge source that can communicate with each subsystem, eliminating the need for constant synchronization and repetitive logic. Knowledge graphs are excellent solutions for solving these data challenges as they can connect the dots between isolated data sources and add semantic context to individual data points.

 

Supporting Scania’s information architecture

One strategic goal for adopting knowledge graphs is to facilitate data sharing across brands within the TRATON group. The knowledge graph is seen as a scalable solution that reduces the need for complex integrations between applications, enabling seamless data sharing and decision-making across the organization.

 

This cross-brand interoperability ensures that all entities within the group can leverage a unified source of truth while minimizing redundancy and effort. By enabling knowledge sharing, we reduce the overall workload on applications, ensuring better efficiency and scalability.

 

Use cases for knowledge graphs at Scania

While the potential applications are vast, some of our key initiatives that leverage the knowledge graph foundation that we have built include:

 

• Purchaser 360: Helps purchasers quickly find alternative parts when supplies run out or are impacted by external events like war or logistical disruptions. The system identifies alternative suppliers or sub-suppliers who can quickly fill gaps, reducing downtime.

• Modernization of Cloud Platforms (MCP): Assesses the complexity of migrating systems to the cloud by mapping the intricate relationships between projects, tables, and schemas. Users can analyze dependencies and determine which projects are critical to migrate first or whether cyclic dependencies exist.

• Autonomous Transport System (ATS): The Knowledge Graph represents the real-time status of autonomous vehicles, factoring in data like road conditions, tire traction, and weather. It applies rules to infer necessary actions, such as slowing down due to wet roads. These inferences are used by fleet managers to monitor and adjust operations, ensuring safe and efficient autonomous driving.

 

Connecting Scania's data to support these use cases with knowledge graphs

A major part of Scania's knowledge graph implementation is about integrating diverse types of data. This includes information about suppliers, geographical data, transportation routes, parts, and design specifications for the Purchaser 360 project. This enables the system to quickly identify alternatives and ensure smoother procurement processes. For the MCP project, the knowledge graph integrates metadata around schemas, tables, and project configurations, helping to understand dependencies and assess the feasibility of cloud migrations. Finally, for the ATS (Autonomous Transport System), the graph is enriched with real-time sensor data from test trucks, providing dynamic insights that drive decision-making related to vehicle performance, safety metrics, and operational efficiency.

 

The importance of reasoning

At the heart of Scania’s knowledge graph efforts is reasoning—the ability to infer new facts automatically based on rules and existing data. This allows us to accelerate decision-making, especially in scenarios where traditionally a human would need to analyze the situation. Whether in the case of autonomous vehicles or purchasing decisions, automated reasoning helps ensure faster, data-driven responses. 

 

Key considerations for building the knowledge graph

Building a knowledge graph at Scania requires careful consideration of several factors, including scalability, ontology design, but also organizational and cultural shifts, as well as management and stakeholder buy-in. 

 

Scalability and Knowledge Graph as a Service

One of the critical aspects we had to consider at Scania was the scalability of the knowledge graph solution. Since Scania is part of the larger TRATON Group, the solution needed to scale across multiple brands, each with its own unique data landscape. This required a scalable architecture where the knowledge graph could be provided as a service (KGaaS), allowing different teams to adopt the technology quickly.

 

By offering KG as a service, we enable users to kickstart their journey with knowledge graphs while still working in familiar ways. Tools like metaphactory have been instrumental in this process. metaphactory allows users to build and interact with knowledge graphs without needing deep technical knowledge. Users can continue creating entity diagrams that describe concepts and relations in the data, much as they did before, but now these diagrams automatically contribute to building ontologies behind the scenes. This empowers users to harness the power of knowledge graphs while maintaining the workflows they are accustomed to, with added benefits like governance and lifecycle management.

 

Ontology design and implementation

Here are the key considerations we took into account when designing and implementing the ontology used in the knowledge graph:

 

• Defining clear competency questions that guide the ontology.

• Involving stakeholders from the outset to ensure the ontology and knowledge graph are aligned with business needs.

• Educating users so they can leverage the platform for their specific use cases.

• Incorporating tools like metaphactory to enable non-technical users to contribute while maintaining governance and lifecycle management over the ontologies being created. 

 

Organizational and cultural shifts

With the adoption of knowledge graphs, we’ve seen greater acceptance now with the Machine Learning (ML) and Artificial Intelligence (AI) teams across Scania. Initially, there was some hesitancy, as teams were unsure how knowledge graphs would fit alongside existing technologies. However, as the benefits became clear—such as how knowledge graphs can enhance data insights and complement AI models—there has been growing enthusiasm from these teams. 

 

This collaboration has opened up new opportunities for integrating semantics with advanced data analytics, driving better decision-making and fostering innovation. Instead of viewing knowledge graphs as competing with existing systems, the teams have started to realize that knowledge graphs and AI work hand in hand, improving the overall ability to deduce new facts and generate better insights.

 

Ensuring management and stakeholder buy-in

Convincing various teams and management that the knowledge graph would add value without competing with existing tools was one of the biggest hurdles. Integrating the graph into Scania’s current architecture and proving its worth required time and persistence. Additionally, ensuring the scalability of the platform across the TRATON group and the ease of use for various departments required strategic tooling and collaborative engagement.

 

To gain support from business stakeholders and domain experts, we have taken a problem-solving approach. By organizing knowledge-sharing sessions, we encourage participants from different business areas to bring their challenges, which we then address using the knowledge graph platform. This results-focused strategy has helped us secure buy-in from a variety of stakeholders. 

 

About the author

Tanuja Gupta

Manager, Knowledge Graphs & Explainable AI

MAN

Tanuja Gupta is a seasoned professional in the tech industry, currently serving as Manager: Knowledge Graphs and Explainable AI at MAN and previously as Solutions Architect and Knowledge Graph Ambassador at Scania (both part of TRATON GROUP). In her role, Tanuja is responsible for designing the KGaas Platform. Tanuja's educational background includes a bachelor's degree in Electronics and Telecommunication and a PG in advanced computing. Her career spans various roles, demonstrating her diverse skillset in technology architecture and project management.