The GBRX AI in Rail Industry Action Plan: A Practical Guide to What Comes Next
5 May 2026
The UK rail industry finally has a clear, practical roadmap for AI adoption.
GBRX’s Artificial Intelligence in Rail: The Industry Action Plan, published in April 2026, sets out exactly how the sector needs to adopt AI to deliver a more reliable, efficient and passenger-focused railway. It’s ambitious, it’s urgent – and for organisations trying to figure out where to start, it can feel overwhelming.
The good news? A lot of what the plan calls for doesn’t require starting from scratch. Netcall’s Liberty platform is already helping rail organisations, from Network Rail to Transport for London to regional train operating companies, like East Midlands Railway, put many of these principles into practice today.
Here’s how Liberty maps to the action plan’s priorities and what that looks like in the real-world.
What the AI in rail action plan is actually asking for
Before diving into the technology, it’s worth being clear on what the GBRX plan is calling for. At its core, it identifies six opportunity areas where AI can make a meaningful difference:
-
Passenger and customer experience
-
Network operations
-
Network planning
-
Rolling stock asset management
-
Infrastructure asset management
-
Organisational processes.
Underpinning all of these are what the plan calls “foundational enablers”, things like data infrastructure, governance, workforce skills, commercial structures and the ability to move from isolated pilots to scalable, reusable solutions.
The plan is very honest about the problem the industry currently faces. Too many promising AI pilots have failed because they couldn’t survive contact with legacy systems, fragmented data or complex multi-party governance. The plan even has a name for it: The pilot graveyard.
The antidote the plan proposes is a pathfinder-led approach: Building solutions in real operational environments, generating reusable components and scaling what works. That’s exactly the philosophy Netcall’s Liberty platform is built around.
Passenger experience: From complexity to clarity
The action plan is blunt about fares: passengers find them too complex and often poor value. It prioritises the integrated fares model and on-journey guidance as key pathfinders; helping passengers navigate disruption, connections and alternatives in real-time, in plain language.
Liberty Converse, Netcall’s AI-powered omnichannel contact centre solution, does this today. East Midlands Railway deployed Liberty Converse in just five weeks, giving passengers self-service options across web and voice while equipping agents with a unified desktop that surfaces customer context immediately. The result was a step-change in how passengers experience disruption — faster answers, clearer information, less frustration.
Liberty also supports accessibility from the ground up. Automated translation, simplified language options and personalised routing mean inclusive design is embedded in the service, not bolted on which aligns directly with the action plan’s emphasis on accessibility and human-centred design.
The plan also highlights the potential of retail agents to help passengers navigate the current fares complexity while wider reform progresses. Liberty’s conversational AI and low-code application tools are well-suited to building exactly this kind of time-bounded, targeted capability — deployed quickly, measured clearly and retired or scaled based on what the evidence shows.
Network operations: Moving from reactive to responsive
One of the action plan’s priority pathfinders is turning free text — Control Centre Incident Logs, maintenance notes, operational reports — into structured, actionable data. Right now, huge amounts of operational intelligence is locked in free text fields that can’t be easily searched, analysed or used to inform decisions.
Liberty IDP (intelligent document processing) tackles this head-on. Using AI to extract, classify and validate information from unstructured documents, it transforms safety reports, engineering records, delay claims and customer correspondence into structured data that can drive workflows and inform decisions. GBRX has already supported work in this area using LLM-based extraction from Control Centre Incident Logs and Liberty provides the platform to scale that kind of capability across the network.
Liberty RPA (robotic process automation) complements this by automating the routine, rule-based tasks that consume operational team time — from delay repay validation to timetable updates and back-office reconciliations. Transport for London used Liberty RPA to migrate over
2,000 hours’ worth of manual data work, safely and without downtime. That’s exactly the kind of middle and back-office automation the action plan describes when it talks about freeing people up to focus on decisions that matter.
For incident management specifically, Liberty Create‘s low-code application development means workflows can be built and adapted quickly without long development queues or heavy IT involvement – think disruption dashboards, automated alerts, mobile-ready tools for frontline teams.
Organisational processes: The quick wins that build confidence
The action plan makes an interesting point that often gets overlooked: Organisational processes across finance, HR, procurement, safety compliance or document management. These are some of the most achievable early opportunities for AI. They don’t interface directly with the live railway, which means solutions can be developed and proven without introducing safety risk.
Network Rail has already used Liberty Create to digitise its RAMS (Risk Assessment and Method Statement) approval process, handling over 600 different request types per month through a single, auditable digital system. What was once a slow, paper-based bottleneck is now a streamlined workflow with automated checks and real-time visibility.
The action plan specifically calls for turning internal free text into structured, actionable insight; processing reports, logs, contracts and correspondence to reduce manual collation. Liberty IDP and Liberty Spark (process intelligence and mapping) are built for exactly this. Spark helps organisations see how workflows, identify where the bottlenecks are and design smarter ways of working. Then Liberty Create or RPA can automate those improved processes.
Foundational enablers: Solving the pilot graveyard problem
Perhaps the most important thing Netcall brings to the GBRX action plan isn’t any single feature, it’s the platform model itself.
One of the plan’s central concerns is that rail AI sits in silos. Solutions are built for one operator, one context, one dataset and they can’t travel. The AIIA (Artificial Intelligence Incubator Accelerator) exists specifically to change this, by creating reusable components, shared patterns and adoption pathways that work across the federated structure of the industry.
Liberty is designed with exactly this in mind. Because it’s a unified, low-code platform rather than a collection of point solutions, capabilities developed for one use case can be adapted and reused elsewhere. The workflow patterns, data models and assurance artefacts from one pathfinder don’t disappear, they become the foundation for the next one.
The plan also stresses the importance of governance and assurance being built into AI adoption from the start, not treated as an afterthought. Liberty’s approach — keeping humans –in-the-loop, maintaining clear audit trails, building compliance into workflows rather than around them — directly supports this. The RAMS digitisation at Network Rail is a good example: The system doesn’t just process approvals faster, it maintains a complete, version-controlled audit trail that meets regulatory expectations.
On workforce and skills, Liberty’s low-code approach means business teams, not just IT specialists, can build and adapt workflows. This is critical in a sector where, as the action plan notes, the industry is losing 5% of its workforce to retirement every year and only replacing them at a rate of 1.5%. Tools that are accessible to non-technical users and that come with support for training and adoption, are essential.
Real rail, real results
Liberty helps complex, safety‑critical organisations escape the pilot graveyard by turning fragmented data, rules and judgement into governed, human‑centred workflows that AI can actually scale.
It’s worth grounding all of this in what’s already happened, not just what’s possible.
Network Rail extended its partnership with Netcall through to 2029, using Liberty Create to build digital workflows across safety, compliance and operational processes including a security check system across managed stations and the RAMS approval platform.
Transport for London used Liberty RPA to safely migrate critical engineering data, saving over 2,000 hours of skilled manual effort, with zero disruption to operations.
East Midlands Railway deployed a full omnichannel contact centre solution in five weeks, with webchat, voice, real-time dashboards and a unified agent desktop, transforming the passenger experience during disruption.
These aren’t proofs of concept. They’re operational deployments, delivering measurable outcomes, in the same complex, safety-critical environment the GBRX action plan is trying to transform.
What this means for train operators and infrastructure providers
The transition to Great British Railways is a once-in-a-generation moment for the industry. The action plan is clear that waiting is not a strategy. AI capability is advancing quickly and organisations that move now will build the data foundations, skills and governance structures that make further adoption easier and faster.
For Train Operating Companies (TOCs) specifically, the pressures are immediate: Rising government scrutiny, performance KPIs under strain, growing digital backlogs and resource constraints that make it hard to innovate while still running the railway.
Liberty’s approach — start with a specific, well-understood problem, build something that works in the real operational environment, generate reusable components and scale — maps directly onto the pathfinder methodology at the heart of the GBRX plan.
The question isn’t whether to adopt AI. It’s whether to do it in a way that creates lasting capability or adds to the pilot graveyard.
Ready to talk?
If you’re a TOC, infrastructure provider or freight operator trying to work out where Liberty fits in your AI journey, or how to make sense of the GBRX action plan for your organisation, we’d love to have that conversation.
About the author
David Oliver
Director of Transport
David leads Netcall's work in the rail sector helping Train Operating Companies navigate one of the biggest transformations the industry has seen in generations. With deep roots in transport and a passion for practical innovation, he works across the industry to modernise operations, cut through digital backlogs and build services that are ready for the Great British Railways era and beyond.