Zooper's route-and-clubbing algorithm and nearest-driver prediction explained in plain language, plus five questions to ask any employee transport vendor that claims AI.
Ask what the AI decides, not whether it exists.
Almost every mobility vendor now says "AI". For a transport head, the useful question is what the AI actually decides. In Zooper, two proprietary algorithms do most of the heavy lifting.
Algorithm 1: route optimisation and employee clubbing
What it decides: which employees share a cab, in what pickup order, and along which route.
What it optimises: the lowest total kilometres driven by all cabs in a shift, and the lowest average travel time for each employee. It works within the constraints your company sets, such as cab capacity, shift windows and night-shift rules.
What it uses: rosters, locations, traffic patterns, road conditions and weather to produce routes and dynamic ETAs. It handles thousands of employees across each shift and improves as it learns from past trips.
Algorithm 2: nearest-driver prediction
What it decides: which driver is likely to be closest to an employee's pickup point for a future ride.
Why it matters: It significantly minimizes empty (dry) runs by dynamically optimizing routes and consolidating passenger drop-offs. That cuts wasted fuel and fleet costs, reduces carbon emissions, and ensures drivers remain productive without unnecessary deadhead mileage.
What AI does not replace
People still set the policy: who rides when, what the night-shift rules are, which vendors to use, and what counts as a delay. The platform executes those rules at scale and flags exceptions. That is the move from execution to monitoring, and it is why configurability matters as much as the algorithm itself.
Why this matters more now
- Scale and shifts. Global capability centres run around the clock across time zones, so a single roster can span several shifts and locations.
- Hybrid attendance. Rosters change weekly, which makes fixed, hand-built routes go stale quickly.
- Cost and carbon. Every kilometre removed helps both budgets and emissions targets.
Five questions to ask any vendor that says "AI"
- What is it optimising: total kilometres, travel time, cost, or all three?
- Which of our policies can it respect, such as seat limits or drop-order rules for night shifts?
- What real-time data does it use?
- How will we measure improvement after go-live (for example dry kilometres, seat occupancy and on-time arrival)?
- What happens on a bad day, such as heavy rain, a strike or a vendor no-show?
We are glad to answer these for Zooper in a demo, with your own roster data if you like. Book a demo.