Chetan Shrouti, Director, AI and Tech Transformation, Transport and Infrastructure, at Klarus, a specialist technology consultancy focusing on AI, on what it will take for Great British Railways to move beyond AI pilots and deliver real operational value.
With Great British Railways (GBR) taking shape as the single, public sector operating body for Britian’s the nation’s railways, the sector has an opportunity to create a more integrated ecosystem by bringing track and train together. However, the real test will be whether, ultimately, these efforts lead to better operational decisions, more reliable services and improved value for passengers and taxpayers.
AI has a crucial role to play in delivering these ambitions but the biggest challenge is the underlying data, governance, interoperability and operating models needed to deploy it successfully. The reality is that organisational integration alone is not enough to overcome fragmentation in data, systems and working practices.
Delivering real operational value, therefore, requires shared digital and operational foundations, trusted information and clear accountability for results. Building these around shared outcomes, from reducing delays and improving maintenance to strengthening passenger confidence, will enable teams across the railway ecosystem to put theory into practice and plan, decide and act together.
Addressing the legacy barriers
The historical issue is that, when tackling disruptions to the network, all too often rail operators rely on manual processes, data scattered across separate documents and siloed technology systems to assess the effect on timetables, rolling stock, crews and passengers. This results in slow and costly recoveries, which fall short of effective, real-time optimisation.
When adopting AI to solve these problems, rail operators face numerous obstacles, including a fragmented ecosystem, operational silos, legacy contracts and ageing assets. All this has created inconsistent data, access rights, duplicate processes and limited interoperability, which make decisions harder to coordinate.
Simply layering AI on top without resolving these underlying barriers would introduce further risk and limit the potential for a successful outcome.
As train operating companies consolidate under DfT Operator Limited (DFTO), the current public owning group, and eventually under GBR, the intended, integrated railway authority, the rail sector has an opportunity to address these legacy barriers and scale AI beyond pilots.
With the Railways Bill progressing through Parliament and integrated regional leadership emerging, now is the time to agree shared operational outcomes and clarify decision responsibilities before fragmented practices become embedded in the new organisation.
Turning strategy into operational delivery
GBRX, the strategic technology body for GBR, launched its Artificial Intelligence in Rail: The Industry Action Plan in April 2026, which created much-needed initial momentum and a crucial coordinated response across the industry. This was an important milestone, helping operators, Network Rail, suppliers and technology partners collaborate around common standards and shared learning.
GBR’s Rolling Stock and Infrastructure strategy, published in September 2026, then set the strategic direction for coordinating decisions on fleet, infrastructure, signalling, electrification, depots and maintenance. However, to turn that overall direction into tangible operational delivery, there needs to be a clear digital and operational foundation.
The strategy groups fleets into four primary categories: intercity, regional express, commuter and metro. Taking this approach as a template, the operational plan could similarly define the technology areas needed to support a more integrated railway into distinct categories. These could include:
- Customer systems, covering journey planning, booking, passenger data and customer service.
- Operational and workforce systems, such as timetables, fleet and crew planning, and HR.
- Engineering and asset systems, including rolling stock, infrastructure and depot management.
- Integration and intelligence, to address data sharing, migration, platform choices and AI-enabled decision support.
The aim should be to enable coordinated decisions across these systems, rather than reproduce existing silos in a new organisational structure.
Turning AI momentum into measurable outcomes
To further capitalise on momentum and deliver tangible results, the coordinated technology plan should firstly define clear goals. Priority use cases need to focus on measurable outcomes that will have the greatest impact, such as reduced delays, lower compensation costs, better maintenance planning, improved safety and increased passenger confidence.
The plan also needs to treat AI deployment as an operating model transformation rather than a technology upgrade. For rail operators, that means starting with the operational decisions that matter the most and then building the data and governance foundations needed to address them.
The AI tools themselves exist. The challenge is having the data quality, operating models and skills to use them successfully and safely within safety-critical networks.
Collaboration also requires exploring data sharing models, which have proved successful in other sectors, such as with open industrial data ecosystems. This would allow the sharing of insights without exposing the underlying data, helping overcome fragmentation while ensuring appropriate privacy and security controls.
Alongside this, assurance needs to be embedded from the outset. Office of Rail and Road’s May 2026 AI plan applies existing regulatory frameworks, while the Rail Safety and Standards Board’s AI Change Toolkit supports assessment of operational change.
These should sit alongside the work on use cases, data and operating models, rather than be treated as a separate consideration at the end of deployment.
Ultimately, AI can only transform the UK rail sector if operators first fix the underlying issues. Bringing track and train together creates a valuable opportunity but success hinges on having the right shared digital and operational foundations.
Those that define the right use cases, strengthen data foundations and work collaboratively with all ecosystem partners will be best placed to scale AI beyond pilots and deliver a safer, more reliable and more efficient railway.

