Interview: Artificial intelligence in the automotive industry

06.05.2026

The introduction of AI in the automotive industry has noticeably picked up speed - initially through the widespread use of the first generative AI tools in the workplace (ChatGPT, Copilot, etc.), and later through a new wave of generative AI (GenAI), large language models (LLM) and AI agents. However, their use is often still uncoordinated: Employees experiment independently, standards and governance are lacking and shadow AI is emerging.
AI Shepherds GmbH is passionately committed to artificial intelligence and supports automotive suppliers and mobility service providers in their AI development. In this interview, CEO Albert Pujol explains why generative AI is the fastest lever for value creation for tier 1 suppliers, how large language models are fundamentally changing quality processes and why responsible governance is now decisive for differentiation, efficiency and the future viability of the automotive industry.

How do you assess the current status of AI implementation at automotive companies and where do you see the most urgent need for action?

Albert: First of all, it is important to point out that we mainly work with Tier 1 suppliers in Germany. My assessment is therefore strongly based on this segment.
At the end of 2024, we observed a significant acceleration in the introduction of AI. Generative AI tools such as ChatGPT, Copilot, Gemini and others have found their way into employees' workplaces on a large scale, allowing a large proportion of the workforce to experience the power of AI first-hand. This momentum was later interrupted by geopolitical and economic factors, such as new US tariffs and the rapid market recovery of Chinese car manufacturers.
By the end of 2025, however, a new wave of AI adoption has emerged. This time it was clearly focused on Generative AI, in particular Large Language Models (LLMs), prompting and AI agents. This is very positive, as these technologies enable enormous increases in productivity at relatively low costs and short implementation times. However, the introduction has not yet been structured or fast enough. Many frontline employees are using AI in isolation and without standards, governance or coordination. This phenomenon is often referred to as shadow AI.
The primary goal must therefore be to coordinate and support the introduction of AI in the best possible way. Specifically, management must actively structure the introduction of AI through clear responsibilities, training programs and standards in order to turn isolated use into sustainable value creation.

Where do you see the greatest value creation potential for AI in the entire automotive industry value chain?

Albert: When we explain AI to automotive professionals, we distinguish between eight AI capabilities: Estimation, classification, computer vision, signal processing, NLP, generative AI, segmentation and pattern recognition. Each of these capabilities adds value in different departments - computer vision is particularly strong in manufacturing, for example, while generative AI has high potential in engineering and administration.
However, the key factor for me is time to value. Most AI capabilities require the development of customized models (mathematical functions), which takes time, data and expertise. Generative AI is fundamentally different. LLMs have already been developed and trained by others and are available to us. This drastically reduces development time and lowers the barriers to entry. As a result, generative AI can be used in almost all departments. In some areas the impact is greater than in others, but overall it represents the fastest and most pragmatic entry into AI for automotive companies today.

What role does AI play in improving quality in the automotive sector and how can large language models accelerate this effect?

Albert: Quality assurance is one of the areas that can benefit most from GenAI. Quality professionals work intensively with documents such as audits, lessons learned, APQP, PPAP, FMEA and 8D reports. By using LLMs, we can now access and reuse historical quality data. Take FMEA creation as an example: a quality engineer could describe the key features of a new product in a chat interface and get suggested failure modes based on similar components from previous projects.
Another strong use case is the use of AI agents in operational quality processes: When they create 8D reports, for example, both customers and suppliers often overlook important information. The quality teams then spend a lot of time on emails and phone calls. AI agents can automatically recognize missing entries, request clarifications and guide users through the process. This saves valuable time and reduces friction in the supply chain.

How does your product DeepQMS help automotive companies to transform their quality processes and achieve measurable results?

Albert: We developed DeepQMS to help quality managers reduce costs, avoid errors and improve performance by automating routine quality tasks with AI. This allows quality experts to focus on standards, prevention and continuous improvement instead of administrative tasks. The current focus is on 8D reports: DeepQMS assists with creation and validation, significantly reducing processing time while improving consistency and quality. So far, we have successfully completed two pilot projects and interest in the solution is growing, including among OEMs. This shows us that AI-driven quality management is moving from the experimental stage to real industrial use.

Thank you very much for the interview and the interesting insights.