From innovation POCs to industrialised solutions. Built by engineers, for engineers.
Weeks → days
Drastic reduction in system requirements allocation time
End-to-end
From business needs and POCs to production-ready AI
At scale
AI deployed across engineering and industrial operations
Putting AI on tracks: Expleo established a dedicated Data and AI service centre to support the innovation division of a global rail leader, turning complex engineering and industrial needs into production-ready AI applications. By combining Expleo’s AI capabilities with deep rolling stock engineering expertise, the AI Factory boosts engineering productivity, accelerates the V-cycle and streamlines access to critical technical knowledge across the organisation.
Who is the client?
Our client is a global leader in rail transport systems, designing and delivering rolling stock, signalling and integrated mobility solutions. With operations spanning more than 60 countries, the group maintains an installed base of 150,000 rail vehicles in commercial service—ranging from high-speed trains and heavy locomotives to metros, trams and turnkey signalling systems.
Designing, certifying and manufacturing these safety-critical platforms involves managing millions of system requirements, complex 3D CAD assemblies and vast volumes of verification and validation (V&V) data across decades-long product lifecycles.
What was the challenge?
The client wanted to scale the use of AI across its organisation, turning a growing range of business needs from different departments into reliable, production-ready solutions. These needs spanned complex, time-consuming activities including tender and requirements management, access to technical knowledge, test reporting and production operations.
The challenge went beyond developing individual AI tools. The client needed a scalable delivery model capable of assessing new use cases, demonstrating their value, developing and industrialising the most promising solutions, and integrating them with existing systems and processes without replacing the legacy IT landscape.
What was the Expleo solution?
Expleo established a scalable AI Factory as a dedicated Data and AI service centre, embedding AI engineers, data scientists, business analysts and solution architects into agile squads. Working hand-in-hand with the client’s engineering teams, Expleo manages the end-to-end lifecycle—from rapid POCs to system integration, deployment and production monitoring.
Pairing 5,000+ Data, AI and software specialists with deep ER&D heritage in mechanical, embedded, RAMS, and manufacturing engineering, Expleo designs every application for immediate field adoption and production readiness.
Combining GenAI, LLMs, RAG, Knowledge Graphs, agentic architectures, machine learning and cloud MLOps, this flexible model industrialises multiple use cases in parallel as programme priorities evolve.
How did this solution help?
The AI Factory has enabled the client to build, industrialise and deploy AI applications tailored to rolling stock engineering and manufacturing:
- Tender and system requirements management: AI automatically parses, classifies and tags thousands of complex requirements from customer tender documents, allocates them to the relevant engineering subsystems (mechanical, traction, brakes, TCMS), and feeds them directly into the client’s DOORS requirements management environment.
- Engineering knowledge assistant: Combining RAG and Knowledge Graphs to give systems and design engineers instant, accurate and traceable access to internal technical standards, legacy project lessons learned and subsystem specifications via a conversational interface.
- Automated V&V test reporting: Extracting and synthesising data from heterogeneous validation sources to automatically generate standardised, compliant test reports based on predefined engineering templates.
- Voice-enabled shop-floor operations: Empowering manufacturing operators and quality inspectors on the assembly line to log non-conformities by voice—automatically generating structured incident records and recommending corrective action plans based on historical shop-floor data.
- 3D CAD similarity search and BOM reuse: Indexing and analysing complex 3D CAD geometries to identify existing qualified components when building BOMs, curbing part proliferation and accelerating mechanical design reuse.
Launched in 2024, the service centre has rapidly expanded as field adoption and internal demand have grown, scaling to eight agile squads working in parallel across an evolving pipeline of engineering and industrial use cases.
What were the results?
The AI Factory is helping our client accelerate processes, reduce manual effort and improve the quality and accessibility of information. AI-assisted requirements allocation can reduce a process that previously took experienced engineers several weeks to just a few days, while knowledge assistants make internal expertise easier to access. On the shop floor, voice-based input makes it easier for operators to report issues directly, without having to type information into an application, while helping capture and reuse lessons learned.
The AI Factory provides a repeatable framework for leveraging AI opportunities. New ideas can be rapidly assessed through POCs, then industrialised when their value and feasibility are demonstrated. Once deployed, solutions are monitored for usage, performance, user feedback and operating costs, including LLM token and resource consumption, providing the foundations to track value and continuously optimize solutions.
The AI Factory is helping our client accelerate processes, reduce manual effort and improve the quality and accessibility of information:
- AI-assisted requirements allocation can reduce a process that previously took experienced engineers several weeks to just a few days
- Knowledge assistants make internal expertise easier to access.
- On the shop floor, voice-based input makes it easier for operators to report issues directly, without having to type information into an application, while helping capture and reuse lessons learned.
Crucially, the AI Factory solves the critical enterprise governance challenge of Shadow AI. By centralising AI development into a single, structured service centre with harmonised methodologies, standardised tech stacks and rigorous LLMOps tracking (monitoring adoption, latency and token consumption), it replaces fragmented initiatives with a scalable, secure framework that is built to expand effortlessly across the organisation.
Could it work for me?
Yes. The AI Factory model is particularly relevant for organisations managing complex engineering or industrial processes, large and scattered knowledge bases, heavy document management processes and significant volumes of technical data. This AI factory model gives organizations a scalable and flexible AI capability, able to respond to evolving priorities while maintaining a strong focus on adoption, cost control and return on investment.
Signalling
We deliver specialised signalling covering all V-model activities, while working with both signalling manufacturers and operators.
Product engineering
From the final design of complex systems to certifications delivered on time and within budget, we make sure products perform as expected in the hands of end users, and we apply AI to streamline the development process.
Contact us to explore our services and solutions
After completing the form one of our experts will get in contact with you as soon as possible.



