GenAI in Motion: Redefining Automotive Systems Engineering - SysEng Copilot

NTT DATA brings GenAI to product development. We have developed a Copilot to support product development in automotive and manufacturing industries. It is based on Microsoft Azure with OpenAI large language models such as GPT-4. As it integrates with existing engineering IT systems such as PTC Codebeamer for requirements and test management, it enhances efficiency, cuts development time, and ensures process conformity.

Index

What is the product vision for SysEng Copilot?
How does SysEng Copilot achieve this vision?
Can you provide a practical example?
What happens next?
What are the benefits of SysEng Copilot?
Can you describe the architecture of SysEng Copilot and how it integrates with existing tools?
How can customers adopt the SysEng Copilot?
Final question: how do you see SysEng Copilot evolving in the future?

Let's hear more from Jens Krueger and Julian Morelli Desanzo. Jens is head of the Global Automotive Engineering Competency and product owner of SysEng Copilot. Julian is Lead Consultant for applied AI in systems engineering with a rich background in AI technology and concepts, he is the solution architect for SysEng Copilot.

What is the product vision for SysEng Copilot?

Jens (Product Owner): SysEng Copilot is streamlining the systems engineering workflow with GenAI, it tackles the complexity of managing processes, methods, and tools (PMT). Along the V model, there are many use cases that can benefit from this approach. Please find more details on our AI use case model for engineering under GenAI in Motion - Redefining the Automotive Product Development Landscape | NTT DATA Group
Although we have this product vision, it is important to understand that this is not a licensable software product. It started as a proof-of-concept and has now become an asset that we use in our projects for clients to show the power of GenAI and to speed-up the implementation of customer-specific solutions.

How does SysEng Copilot achieve this vision?

Julian (Systems Architect): SysEng Copilot is built around two main components: Method Finder and Method Coach. The Method Finder intelligently recommends the best-fit processes, methods, and tools from a PMT architecture database. Once a selection is made, the Method Coach provides step-by-step guidance and automates interactions with engineering IT tools via their APIs.

Can you provide a practical example?

Julian: Certainly. Please also refer to the following video. Imagine you're at the stage of generating test cases from system requirements in a car development project. By querying SysEng Copilot, the Method Finder would recommend the "Automated Test Case Generation from System Requirements" method, tailored for automotive development projects and integrated with tools such as PTC Codebeamer.

What happens next?

Julian: Once the method is selected, the Method Coach takes over, providing detailed guidance and automating tasks. For example, it can map system requirements to test scenarios using predefined templates, ensuring comprehensive test coverage and alignment with industry standards such as Automotive SPICE and ISO 15288 Systems Engineering.

NTT DATA SysEng Copilot demo（3:28）

What are the benefits of SysEng Copilot?

Jens: First of all, it is a great starting point for GenAI-based solutions in product development. With our comprehensive use case model and the solution architecture template based on Microsoft Azure, our customers achieve very fast time to value. This value is for example enhanced efficiency by automating workflow steps, optimized resource usage by recommending appropriate tools, and improved process conformity by tailoring methods to company standards.

Can you describe the architecture of SysEng Copilot and how it integrates with existing tools?

Julian: SysEng Copilot is built on Microsoft Azure using Retrieval-Augmented Generation (RAG) and AI agents. RAG is a technique that enhances the accuracy and reliability of generative AI models by fetching facts from external sources, filling a gap in how large language models (LLMs) work. AI agents are software programs that can act autonomously in an environment, transforming data and making decisions. The architecture of SysEng Copilot is structured into four layers: User Interface, AI & Data Processing, Data Storage, and Application Integration. It also integrates with external applications like IBM DOORS, EA, Confluence and Jira via API. This enables the actual automation of method steps, for example the generation of test cases for a given requirement in PTC Codebeamer.

How can customers adopt the SysEng Copilot?

Jens: Just contact us! We are offering a 3-month Value Discovery PoC together with our partner Microsoft. We start by identifying the use cases with the best combination of business impact, feasibility and effort. These use cases are then implemented in an Azure subscription of the customer, so that the data stays there. Finally, we evaluate the use cases as a team and provide recommendations including a roadmap for scaling the PoC solutions.

Final question: how do you see SysEng Copilot evolving in the future?

Julian: I see it evolving fast! We have implemented the initial PoC in only two months and this field is changing so quickly. So, we will continuously integrate the best AI technology and solution patterns into our architecture template.
Jens: A strong architecture is the basis for innovative use cases. Over time, I am looking forward to having our comprehensive use case model for AI in engineering implemented in an integrated way, supporting systems engineering processes and methods of our customers.

Jens Krueger

Head of Global Automotive Engineering Competency, NTT DATA Germany

Julian Morelli Desanzo

Lead Consultant Applied AI in Systems Engineering, NTT DATA Germany

Contact

NTT DATA Group Corporation
Global Marketing & Communication Headquarters
Shinsuke Yoshinaga, Ayaka Matsuzawa, Toshiki Takei
E-Mail: global-marketing@kits.nttdata.co.jp