The automotive industry is at a crossroads. Rising costs, mounting sustainability demands, stringent financing conditions, and the complexities of regionalization amid escalating geopolitical tensions pose significant barriers to growth. At the same time, advanced innovative technologies like generative artificial intelligence (AI) present a transformative opportunity, especially for research and development (R&D) and engineering functions.

Companies that deploy generative AI strategically can boost efficiency and respond quickly to accelerated development cycles and the growing complexity of connected vehicles. Beyond testing and prioritizing use cases, companies must succeed in deploying use cases at scale if they want to see the profit and loss (P&L) impact of AI. We believe seven key principles are instrumental to successfully scaling AI and staying ahead of the competition.

1. Establish a transformative long-term vision for AI integration

Successful AI transformation requires a clear long-term strategic vision, bridging the gap between proof-of-concept and large-scale implementation. Companies must identify long-term disruptions and link them to short-term quick-win initiatives. This approach will form coherent chains of use cases, unlock broader value pools, and provide a clear roadmap forward.

For R&D specifically, we anticipate major disruptions in the design and development and testing and validation phases, thanks to technologies like generative design, generative computer-aided design (CAD), virtual simulation, and advanced knowledge management. For instance, the R&D executives we worked with anticipate powerful tools leveraging libraries of generic products to design and test simultaneously, hence with the potential to disrupt the traditional development process.

2. Ensure AI-readiness of product lifecycle management

A robust product lifecycle management (PLM) system is instrumental to any digital transformation in R&D. PLM must ensure digital continuity and interoperability across systems, and it is estimated that more than 70% of PLM applications will support the end-to-end digital threads strategy by 2027.

An AI-ready PLM is a critical enabler for integrating and deploying at scale advanced AI applications seamlessly. There are several dimensions to check when deploying a new PLM solution that is AI-ready: cost efficiency, performance, customizability, upgradability, UI/UX, data management, collaborative features, vendor support, and the choice between on-premises and cloud options.

3. Upgrade the maturity of your data for a robust AI strategy

If AI is the engine driving R&D efficiency, then data is the fuel. A robust AI strategy must be supported by an equally strong data strategy, focusing on both volume and AI maturity. For R&D, capturing large volumes of data is crucial to improve learning and outcomes, while ensuring that data is easily accessible, high quality, and interconnected is instrumental to enabling AI use cases.

Ensuring quality and accessibility is crucial in R&D because the data involved is often complex or unstructured, such as 3D CAD files. Additionally, this data can be less accessible to AI solutions, especially when it is stored on individual computers.

4. Build a modular architecture with a clear make-or-buy strategy

To successfully integrate AI in R&D and engineering, companies must establish a modular AI architecture that allows for scalability, flexibility, and upgradability in a context of a fast-moving technology landscape. This means adapting AI agents/models to each use case, with an objective of infrastructure frugality, and industrializing development pipelines with reusable components and repeatable processes.

A clear make-or-buy strategy becomes crucial: Organizations will usually build or customize AI solutions internally for differentiating use cases and when company-specific knowledge is at stake. Off-the-shelf AI components are often used for more transactional use cases.

Reliance on vendors must be carefully managed, particularly in terms of cost. Our analysis of AI companies’ data suggests that an estimated 10% of costs per query is passed to users. Ultimately, a well-designed and maintained AI ecosystem ensures access to diverse expertise, accelerates development, and keeps alignment with business needs while avoiding the common “Not Invented Here” pitfall in R&D.

5. Set up a hybrid organization model with C-suite sponsorship to drive AI impact

AI initiatives require more than technical expertise. Organizations also need effective cross-functional governance at the C-suite level to drive adoption and scalability. This is particularly vital for R&D and engineering due to the numerous interfaces with other functions, which necessitate close oversight to prevent redundant efforts and maximise synergies. Companies need to balance strategic central oversight with local innovation, allowing teams to experiment with new use cases and technology while avoiding excessive cash burning.

Governance also plays a critical role in monitoring risks associated with AI development, including ethics, cybersecurity, and compliance. Particularly in R&D, safeguarding business-critical data and intellectual property is essential, which may lead to strategic decisions like relocating data-rich activities.

6. Address data sovereignty in a global context

AI stands at the intersection of various geopolitical and regulatory dynamics. It is essential for organizations to safeguard their sovereignty in relation to specific regions/nations. In the context of heightened regionalization, data is becoming an increasingly sensitive asset, and organizations must safeguard their data sovereignty. This involves:

• Defining a clear strategy for data protection and exchange, as high-quality, proprietary data serves as a crucial raw material for AI, enabling organizations to gain unique insights and differentiate themselves in the market
• Selecting the right technology vendors for each region with targeted local strategies; regulation on AI usage is emerging but is far from uniform across regions, especially in the European Union, China, and the US
• Securing business continuity to anticipate disruptions in data systems due to regional blockages that can jeopardize overall operations

7. Anticipate the impacts of AI on the research and development workforce

The impact of AI on the workforce will be significant, particularly in the design and development testing and validation phases. For most impacted jobs, we estimate that efficiency gains will range between 10% and 35%. Overall, the average impact across jobs and R&D phases is expected to reach 15% to 20%. But to achieve those new levels of efficiency, organizations must create a culture around AI transformation, including bolstering staff training on such topics as data awareness, innovation, agility, and collaboration. Continuous learning, reskilling, and upskilling are essential to empower employees to leverage AI technologies and contribute to ongoing improvements and solutions.

How to make sure your company is AI-ready

The automotive industry stands at a pivotal moment. With AI and generative AI expected to disrupt and redefine R&D processes, companies that proactively build and steer their AI strategic vision, ensure their technological and data readiness, and anticipate AI impact on their workforce will be best positioned to lead the market. The challenge is not only to explore the myriads of AI use cases but to deploy them at scale to materialize efficiency gains in the P&L and secure competitive advantage.