
Embedded with operational rail teams to build vacancy, roster and train-planning products, moving critical workflows from fragmented spreadsheets and slow validation loops into a connected, governed environment.
A national rail-freight operator whose planning processes spanned timetables, diagrams, rosters, crew skills and vacancies.
Critical planning processes ran across specialist source systems, spreadsheets, email and inherited Foundry applications. Timetable changes, diagramming, rosters and crew vacancies were closely related operationally but difficult to reconcile digitally. Slow validation cycles and inconsistent source data added noise, while previous unsuccessful technology programmes had created understandable scepticism among users.
Quant8 embedded engineers with planners, rostering specialists, control teams and technology colleagues, developing a connected operational capability across three workstreams rather than one application.
Vacancy management:
Base-roster planning:
Train planning and grids:
The programme used staged adoption rather than a single handover: internal functional testing, subject-matter-expert testing, broader team testing, formal sign-off, training material, go-live and hypercare. Control-room users were trained during rollout, with further location-based training planned, and user feedback was incorporated directly into working software, including changes made during live sessions.
The vacancy product completed production rollout and hypercare. The base-roster product had passed almost all test items and was moving towards production sign-off. The connected planning work had progressed from disconnected manual grids to a working flow builder and governed data foundation, with broader user demonstrations underway.
The engagement also identified a material optimisation opportunity in an inherited pipeline: removing unnecessary processing and data round-tripping is projected to reduce a slow one-to-two-hour validation cycle to minutes and create significant compute headroom. This remains an identified, projected benefit rather than a realised saving.
A proudpartner, delivering enterprise AI at scale.
Bring us a specific operational problem and we'll show you how we would approach it.