The situation
The operations team at a Singapore-based commercial vehicle fleet company was managing 120 vehicles across multiple client contracts. Route planning was done manually each morning by three senior dispatchers, each spending two to three hours building daily schedules in spreadsheets and cross-referencing against driver availability, vehicle maintenance windows, and client delivery windows.
Beyond planning, the team was drowning in administrative overhead: driver briefing documents, client delivery reports, fuel reconciliation, and compliance paperwork for LTA requirements. The operations manager estimated that 60% of her team's day was spent on tasks that required no judgement — just data entry, formatting, and distribution.
The company had explored off-the-shelf fleet management software but found the implementation costs prohibitive and the tools too rigid for their mix of contract types. They needed AI that worked within their existing systems, not a platform replacement.
What we did
- 01
Spent the first week embedded with the dispatch team, riding along on the morning planning process and documenting every step. Identified that route planning alone was consuming 63 dispatcher-hours per week — the single largest time sink in the operation.
- 02
Redesigned the route planning workflow to use AI-assisted optimisation within the team's existing tools, eliminating the manual spreadsheet process while preserving dispatcher control over final decisions. Dispatchers went from building routes from scratch to reviewing and approving AI-generated schedules.
- 03
Automated the generation of driver briefing documents, client delivery reports, and fuel reconciliation summaries — three high-volume, low-judgement tasks that were consuming an estimated 20 hours per week across the team.
- 04
Ran two structured training sessions with all 18 operations staff, including a dispatcher-specific playbook. Tracked adoption and time savings daily for the first three weeks post-deployment, with weekly check-ins for the following month.