Project Summary
Trackmate is an AI-powered platform that allows passengers to ask questions like ‘How do I get to Edinburgh tomorrow?’ either verbally or via text and receive comprehensive journey plans. The platform integrates data from National Rail and operators of services including buses and trams into a single intelligent interface, enabling users to plan journeys, access real-time travel information, and purchase train tickets which are automatically added to their Google or Apple Wallet on their phone.? , ? The platform eliminates the need to navigate multiple apps and websites, making public transport more accessible for elderly users, people with disabilities, tourists and anyone seeking simpler travel planning and seamless ticket purchasing.
Project Achievements
“Pocket is a production-ready, AI-powered multi-modal transport platform that reimagines how passengers interact with public transit. Pocket makes the interaction genuinely simple, accessible, and human. We built a voice-first interface using ElevenLabs speech technology for hands-free, accessible journey planning, and an AI Avatar assistant deployed in ticket offices to support passengers during disruptions, reducing pressure on staff exactly when it matters most. The platform spans web, mobile, WhatsApp, and static kiosks, meeting passengers wherever they are. We also delivered multilingual support across English, Spanish, Catalan, and Welsh, with an architecture that’s built to grow. Under the hood, a document intelligence system using vector search gives passengers instant answers from live policy and timetable data. Pocket doesn’t just automate transit information; it makes public transport feel like it was designed for people. “
Conclusions
Pocket confirmed that natural language removes the complexity barrier that has always made public transport intimidating. Passengers shouldn’t need to memorise station codes, decode fare structures, or juggle multiple apps. They shouldn’t even need to speak English. LLMs proved highly effective at handling the ambiguity inherent in real passenger queries, interpreting intent rather than requiring exact inputs. This is the shift that makes conversational AI genuinely useful in transit, not just technically novel. Integrating live bus, rail, and metro data into a single coherent experience is hard. The data sources are messy, inconsistent, and siloed. But the project proved that the complexity is worth absorbing on the backend. Conversations with industry experts from TfL, TfW, a Spanish transport operator, and investors pointed to a clear sweet spot – incident management. Most passengers know their daily route. But when signal failures or trespassers bring a station to a standstill, the information office is overwhelmed, and ticket office queues snake down the platform. Pocket is built exactly for that moment. Pocket absorbs the pressure, giving passengers the answers they need in their own language, and letting staff focus where humans are needed most.
Next Steps
Pocket is now focused on scaling for commercial deployment. We are actively seeking seed investment to grow the engineering team, fund a dedicated sales resource, and accelerate pilot programmes with UK transit operators. Commercially, our primary targets are Transport for Wales, Transport for London, and regional multi-modal operators such as Manchester’s Bee Network. We are also pursuing white-label partnerships with existing transit app providers, positioning Pocket as a high-value bolt-on to operator platforms. On the product side, we plan to expand coverage to include coach and ferry services, integrate open ticketing standards for seamless multi-operator fares, and add predictive disruption alerts using historical pattern analysis. Internationally, Pocket is well-suited to tourist-heavy cities where language is a barrier to using public transport. We have already opened conversations with a Spanish travel operator to pilot Pocket for international visitors navigating local networks.
