For most organisations, internal generative AI productivity begins with drafting, summarising and enterprise search. For an automotive manufacturer, the opportunity — and the risk — is materially different. AI is being introduced into the work that defines, tests and validates products that must remain safe, compliant and maintainable for years after they leave the factory.
in an interview published by Volkswagen, Werner Tietz, Head of Group Research and Development, Volkswagen Group stated:
“At Volkswagen, AI is not an add-on or a standalone project – it is an integral part of how we work.”
Volkswagen Group says it has more than 1,200 active AI applications across the organisation, with hundreds more in development or approaching implementation. In Technical Development alone, it says more than 100 AI-based processes entered productive use during 2025. The programme spans design, simulation, software testing, validation, production quality, supply chain work, cybersecurity and internal knowledge exchange.
That scale makes Volkswagen’s approach relevant beyond automotive. It is an example of how a large industrial business is attempting to turn generative AI from an employee-assistance tool into an engineering-system capability. The prize is substantial: faster product decisions, reduced documentation effort, better reuse of engineering knowledge and shorter routes to series production. The constraints are equally substantial: complex supplier ecosystems, intellectual property, legacy systems, regulated safety processes and the need for engineers to remain accountable for every final decision.
Related Articles
- Extended Reality Automotive Use Cases: Real Ways to Use XR in the Automotive Industry
- Ford’s Big Bet on Microsoft Teams – and the Risks Behind a 170,000-User Migration
TL;DR — What Automotive and Industrial Leaders Should Know
- Volkswagen says more than 1,200 AI applications are active across the Group, spanning engineering, production, supply chain, cybersecurity and administrative work.
- More than 100 AI processes entered productive use in Technical Development in 2025, according to Volkswagen.
- The Group’s GHOST tool automates infotainment software testing, with Volkswagen reporting reproducible tests, fewer documentation errors and faster release cycles.
- Volkswagen is targeting vehicle-development time of under 36 months, around 25% faster than its current baseline. This is a stated ambition, not a published achieved outcome.
- The productivity case depends on controlled implementation: AI can accelerate analysis and documentation, but engineers retain responsibility for system decisions and validation.
Why Automotive Engineering Needs a Different AI Productivity Model
Direct answer: Automotive engineering creates a far more demanding AI productivity environment than general knowledge work. Every requirement, test case, software change and supplier input may influence a safety-critical product. The most useful AI tools are therefore those that reduce repetitive engineering work while preserving traceability, review and human accountability.
Modern vehicles are no longer primarily mechanical products. They are complex systems of software, hardware, sensors, electronics, cloud services and digital interfaces. That means the work required to define and validate a vehicle increasingly resembles enterprise software delivery — but with the additional burden of functional safety, compliance, physical-world testing and long product lifecycles.
Volkswagen’s strategic response is to apply AI throughout that lifecycle rather than confining it to isolated pilots. The Group says AI is being used from early design and material construction through testing, simulation and validation. Its stated goal is to bring development time below 36 months, which it describes as approximately 25% faster than today.
The distinction is important for productivity buyers in manufacturing, aerospace, energy and other industrial sectors. A generic copilot can speed up drafting and retrieval. A sector-specific AI productivity programme must improve the flow of work between requirements, design, testing, evidence and approval — without breaking the audit trail that makes complex engineering governable.
Inside Volkswagen’s Engineering AI Use Cases
Volkswagen’s clearest public example is GHOST, a proprietary tool that automatically tests software functions in infotainment systems, including touch interactions. The Group says the system effectively “presses buttons” like a human and delivers reproducible tests, no documentation errors and faster release cycles.
The benefit is not simply automated test execution. It is the reduction of a familiar engineering bottleneck: skilled staff spending time documenting, repeating and reconciling tests instead of investigating edge cases, refining functions and solving higher-value problems. In that sense, GHOST represents a more credible industrial AI use case than a broad promise that “AI makes engineers more productive.”
Volkswagen is also working with PTC and Microsoft to integrate Microsoft Copilot capabilities into PTC Codebeamer, an application lifecycle management platform used to define, validate and trace requirements and tests across hardware and software development.
In a Microsoft customer story, Robert Kattner, Head of Volkswagen Group IT Engineering said:
“By having a copilot in the Codebeamer software, it can assist with creating new requirement specifications and test cases using our specific data and business context.”
The platform is intended to help engineers search, analyse and generate content within the lifecycle-management environment, import requirements from legacy systems, and support workflows for authoring, reviewing and validating requirements. PTC says Codebeamer customers can realise up to 20% to 40% time savings in these data and requirements workflows. That is a vendor claim rather than a Volkswagen-specific published result, but it indicates the type of productivity opportunity being targeted.
Does the Evidence Support Volkswagen’s AI Productivity Case?
Direct answer: Volkswagen has supplied credible evidence of deployment scale and tangible engineering use cases. It has not yet published enough independent, task-level outcome data to prove that its wider 25% vehicle-development acceleration target has been achieved.
There is strong evidence that Volkswagen’s programme has moved beyond experimentation. The Group reports active applications at scale, productive technical-development processes and specific AI tools in use. It has also invested in skills, launching its WE & AI training and dialogue initiative in spring 2024, and establishing a professorship in AI Methods in Product Development with TU Braunschweig.
What remains less clear is the aggregate productivity impact. Volkswagen’s claims around fewer documentation errors and faster release cycles are plausible for controlled testing workflows, but public evidence does not yet quantify their effect on quality, defect escape rates, engineering headcount capacity or time to market across the Group’s multi-brand portfolio.
That does not invalidate the strategy. It defines the next proof point. As enterprises move from individual AI applications to AI-enabled operating models, they need to measure whether local improvements compound into meaningful product-development gains.
Human Accountability Is Still the Industrial Differentiator
Volkswagen’s public position is that AI should enhance engineering judgement rather than replace it. The Group states that human judgement will always make final decisions in sensitive personnel matters, and that engineers remain responsible for the systems they develop.
Tietz later added:
“We often refer to ‘AI-accelerated engineers’ – professionals who see AI not as a substitute, but as an amplifier of their capabilities.”
That framing is not merely cultural. In automotive development, AI-generated requirements, test cases or analysis can accelerate work, but they cannot become the final authority. A model may produce a convincing output that omits a safety condition, misunderstands an interface dependency or replicates flawed source data. The operational model must therefore preserve human review, version control, traceability and a clear record of what the AI contributed.
Volkswagen’s software-defined vehicle strategy adds urgency. The Group is pursuing scalable electronic architectures, over-the-air software updates and more integrated development between software, data and hardware. Its China programme is important to that strategy, but it is only one part of a broader global shift: Volkswagen operates 111 production facilities across Europe, the Americas, Asia and Africa, and sells vehicles in more than 150 countries.
The relevant lesson for global manufacturers is not to copy one market’s technology roadmap. It is to build a common productivity foundation that can support local development, regional regulation, supplier collaboration and varying customer expectations without fragmenting the engineering operating model.
Verdict: Volkswagen Has Moved Beyond AI Experimentation
Volkswagen’s internal GenAI productivity story is compelling because it is anchored in real industrial work. The Group is applying AI to testing, requirements management, simulation, documentation and validation — processes where time savings can be valuable, but only when engineering control remains intact.
The evidence supports Volkswagen’s claim that AI is becoming embedded across its development organisation. It is less conclusive on the Group-wide business impact of that transformation. The sub-36-month development goal is ambitious and will require more than AI tools: it depends on software architecture, supply-chain collaboration, data quality, modern lifecycle management and disciplined change management.
Buyer verdict: For automotive and industrial enterprises, Volkswagen offers a useful model for AI productivity: begin with repeatable engineering friction, integrate AI into trusted development systems, retain human accountability and measure outcomes against release speed, traceability and quality. AI can accelerate engineering. It cannot carry the engineering responsibility.
Frequently Asked Questions: Volkswagen AI Productivity and Automotive Engineering
How is Volkswagen using AI in automotive engineering?
Volkswagen says it uses AI across vehicle development, software testing, simulation, validation, production, supply chain, cybersecurity and administrative work. Specific engineering applications include its GHOST infotainment testing tool and AI-supported requirements management in PTC Codebeamer with Microsoft Copilot capabilities.
How many AI applications does Volkswagen have?
Volkswagen Group says more than 1,200 AI applications are active across the Group, with hundreds more in development or nearing implementation. Its Head of Group Research and Development has also said more than 100 AI-based processes entered productive use in Technical Development during 2025.
What is Volkswagen GHOST?
GHOST is Volkswagen's proprietary AI-supported tool for automatically testing infotainment software functions, including touch interactions. Volkswagen says it improves test reproducibility, reduces documentation errors and supports faster release cycles.
What is Volkswagen's AI development-time target?
Volkswagen has stated an ambition to reduce vehicle development time to under 36 months, which it describes as around 25% faster than its current baseline. This is a forward-looking target rather than a published achieved outcome.
How does Volkswagen use Microsoft Copilot in engineering?
Volkswagen is working with Microsoft and PTC to bring Copilot capabilities into PTC Codebeamer. The aim is to help engineers create requirement specifications and test cases, search engineering content, analyse information and support requirement review and validation workflows using Volkswagen-specific data and context.
Can AI replace automotive engineers?
Volkswagen's stated approach is that AI should amplify engineers rather than replace them. AI can take on repetitive testing, analysis and documentation tasks, while engineers retain responsibility for ideas, system decisions, validation and final accountability.
About the Author
Alex Cole is a Technology Journalist and Videographer at UC Today. He has experience reporting on productivity and automation, human capital management and extended reality. Connect with Alex on LinkedIn.
