David Oliver, Transport Sector Director, Netcall, on why strengthening operational processes, data and digital foundations should come before large-scale AI deployment across Great British Railways.
As AI usage grows, organisations across the rail industry are under growing pressure to show how they are preparing for a more intelligent, customer-focused future – from operational planning through to passenger services. However, as plans surrounding Great British Railways (GBR) continue to accelerate, there is a temptation to prioritise AI innovation before putting the operational foundations in place to support it.
We already know AI has a key role to play in the future of rail. The crucial step now is to figure out whether the systems, processes and data underneath it are robust enough to make AI genuinely effective at scale.
Many operational processes across the rail network continue to rely on disconnected systems, manual intervention and inconsistent ways of working. While these challenges may appear relatively small in isolation, they collectively limit AI’s ability to deliver value at scale.
AI is only as good as the operation behind it
As operators and infrastructure move closer together under GBR, there will inevitably be a period of consolidation and adjustment across the industry. Different operators have developed their own systems, processes and operational approaches over many years. Bringing those environments together successfully will require far more than simply introducing new AI tools.
The reality is that AI is only ever as effective as the data and operational structure supporting it. Poor-quality data simply creates poor-quality outputs, regardless of how advanced the technology itself appears. That is why there is growing caution across the sector around rushing into large-scale AI projects without first addressing the smaller operational gaps that exist underneath them.
Many of those gaps are not particularly visible from the outside. They are everyday operational processes, incident management workflows, access requests and reporting structures that still rely heavily on manual intervention or disconnected systems. Yet these are exactly the kinds of processes that generate the trusted operational data needed for AI to work effectively.
There is also a growing recognition that some AI projects across the industry are struggling because the underlying operational foundations are not yet mature enough to support them. Rail organisations understandably want to innovate, but there is a difference between piloting AI and deploying it effectively at scale across a fragmented operational environment.
Digitisation before transformation
This is something Network Rail has already been working to address through the digitisation of operational and exception management processes across parts of the network. Rather than focusing purely on large-scale AI ambitions, the approach has been centred around improving operational consistency, standardising workflows and creating better data integrity across systems.
One example is Network Rail’s work to digitise land and consent management processes using low-code and workflow automation technology. By bringing workflows, approvals and specialist support into a more connected digital environment, the organisation has improved visibility across projects, reduced delays caused by manual handling and created a more consistent operational process overall.
Network Rail has also digitised parts of its RAMS (Risk Assessment Method Statement) approval process, replacing heavily manual workflows with faster and more transparent digital approvals. By improving visibility across teams and reducing administrative burden, projects like these help create more structured operational data across the wider network.
While projects like these may not immediately be described as ‘AI transformation’, they are exactly the kind of operational improvements that help create cleaner workflows and more reliable data environments for AI in the future.
Modernising without starting over
The transition towards GBR is a long-term programme and most rail organisations cannot afford to replace entire estates of legacy technology overnight. That is why many are moving towards more flexible, platform-based approaches to transformation and technology investment.
Rather than replacing every existing system, the focus is on connecting processes and operational data more effectively across the network. The aim is to reduce duplication, improve visibility and create a more reliable picture of what is happening across operations without forcing organisations to start from scratch.
Technologies such as low-code development, workflow automation, robotic process automation (RPA) and system integration are helping organisations modernise gradually while continuing to work with existing systems. In practice, that can mean automating repetitive operational tasks, digitising manual approval processes, connecting disconnected data and giving operational teams faster access to accurate information.
A platform-based approach also allows organisations to introduce change incrementally, while creating more joined-up ways of working across operators and infrastructure over time.
The direction of travel across the industry is already clear. The GBRX AI Industry Action Plan highlights the importance of better data-sharing, stronger digital foundations and greater collaboration across rail organisations if AI is going to deliver meaningful value at scale.
Importantly, technologies like workflow automation and RPA are already helping rail organisations remove many of the repetitive administrative tasks that slow operational teams down today. AI will increasingly build on top of those foundations, but the immediate value often comes from improving visibility, consistency and operational efficiency first.
AI should enhance operations, not distract from them
Innovation across the technology sector is accelerating and the rail industry must keep pace. But equally, the industry cannot lose sight of operational delivery in the pursuit of AI innovation.
The organisations best placed to succeed over the next decade are unlikely to be those making the boldest AI announcements today. They will be those investing in stronger operations, trusted data and technology that can evolve alongside the wider transformation of GBR.
That approach may feel less ambitious than some of the AI promises dominating the conversation. But in a complex industry like rail, meaningful transformation rarely happens through one giant leap. It comes from making steady progress and getting the fundamentals right.
