Japanese technology firm Fujitsu has announced a strategic consulting partnership with UK-based Rinicom Ltd to bring Yellowline AI, an early-stage rail safety system combining computer vision and large language models, to market across the UK rail network.
According to Fujitsu, Yellowline AI analyses passenger behaviour on rail platforms in real time and generates personalised, situationally aware public announcements designed to reduce unsafe platform behaviours and prevent passenger harm. A pilot demonstration showed detection accuracy of 90% or above with latency of two seconds or less.
Fujitsu's consulting engagement covered business case development, market research and competitive analysis, go-to-market strategy, stakeholder mapping and business model innovation, including a move towards an AI-as-a-Service licensing model to support scalable deployment across rail operators.
The collaboration culminated in a successful minimum viable product and proof of concept, identifying market entry pathways and laying the groundwork for further development under Fujitsu's Vision AI strategy for the transport sector.
Yellowline AI was recognised as a winner in the UK Department for Transport's First of a Kind 2025 rail competition, run in partnership with Innovate UK, providing further validation of its potential within the UK rail safety market.
"This impactful collaboration goes beyond consultancy," said Siddharth Das, managing consultant at Fujitsu. "It's about deeply embedding our expertise in product strategy and go-to-market advisory to accelerate a truly groundbreaking rail safety solution."
Garik Markarian, chief executive of Rinicom, said Fujitsu's guidance had validated the company's vision and provided a clear roadmap for bringing Yellowline AI to the UK market.
The system has the potential to reduce delay penalties, limit asset damage and decrease the number of incidents per year per station across the rail network.
Access the full announcement to explore the Yellowline AI technology and Fujitsu's rail safety strategy in detail.




.png)
