Wildberries

Autoqueue — Automated Dispatch & Human Control

Warehouse vehicle queues with automatic dispatch, manual intervention, and a record of changes.

Route queues with operating modes and departure targets. Operational data has been anonymized

Outcome Released warehouse dispatch workflow

Role: Senior Product Designer
Team: Product manager, 2 business analysts, engineers, QA
Scope: Route queues, manual control, pinning, departure targets, permissions, event history

As the sole designer, I designed the dispatcher workflow for managing vehicles waiting to load or unload at Wildberries warehouses.

The full workflow was released. Dispatchers can pin a vehicle in place, return unpinned vehicles to automatic control, and review who changed the queue.

The dispatcher’s task

Dispatchers decide which vehicle can proceed to a warehouse gate for loading or unloading. They need to account for vehicle readiness, trip type, compatible gates, and available capacity.

Previously, dispatchers checked several sources, contacted drivers, and coordinated exceptions in chat. A vehicle’s position could change without a visible explanation, and manual adjustments were difficult to trace. This case focuses on the dispatcher’s queue and control flows.

Why manage the queue by route?

One vehicle can be compatible with several gates. Separate lists for each gate split the information needed to decide which vehicle should move next. I used the route as the shared view of its vehicles, gates, and trip types, with a detailed queue for individual actions.

Readiness, departure targets, trip type, and pinned vehicles affect queue order. I placed explanations beside the priority controls so dispatchers could check those factors before intervening.

Detailed route queue with vehicles, trip types, statuses, and driver contacts — A route’s queue shows vehicle position, trip type, status, and driver contacts

Explanations for departure-priority controls, quotas, and vehicle pinning — Explanations beside the controls show what affects priority

Example: hold one vehicle in place

In this illustrative scenario, a dispatcher needs to keep one vehicle in its current position. They start a time-limited manual session and pin the vehicle. Its position is preserved during automatic recalculation, while the rest of the queue can keep moving.

I designed pinning for these local exceptions: the operator can preserve a specific decision without replacing automatic control of the whole queue.

Route queue with manual control enabled for dispatcher intervention — Start manual control, then pin the vehicle whose position must be preserved

Pinned vehicle that remains fixed during automatic queue recalculation — Start manual control, then pin the vehicle whose position must be preserved

Return the queue to automatic control

When the dispatcher turns manual control off, unpinned vehicles may change position as the algorithm recalculates the queue. I added a confirmation that explains this consequence before the change, followed by a state showing that automatic control has resumed.

Confirmation explaining what happens to unpinned vehicles when manual control is disabled — The confirmation explains what happens to unpinned vehicles when manual control ends

Route queue after manual control is disabled and automatic movement resumes — The confirmation explains what happens to unpinned vehicles when manual control ends

Set the mix of departures

Arrival order alone could not maintain the required mix of direct and round-trip departures. I designed a form for setting that ratio, with validation when both values are zero. After saving, dispatchers can compare the target with actual departures.

Departure-ratio validation when both trip-type values are zero — Reject a zero-to-zero ratio, then show the saved target alongside actual departures

Route queue after the departure ratio has been saved — Reject a zero-to-zero ratio, then show the saved target alongside actual departures

Trace a queue change

To review a manual adjustment, a dispatcher can open the vehicle’s history. I structured it around status changes, calls, confirmations, position adjustments, and their initiators. Event filters help operators find the relevant change and see who made it and when.

Vehicle history with status changes, calls, confirmations, queue adjustments, and initiators — Vehicle history records the event, its time, and who initiated it

My role and collaboration

My design work covered queue views, manual-control flows, validation, and event history. I worked with a product manager, two business analysts, engineers, and QA to align permissions, queue states, concurrent editing, and the return to automatic control. The handoff documented session limits, pinning behaviour, validation rules, and event types.

Impact

Post-launch waiting-time and adoption figures are not included in this case. Follow-up measures would include vehicle waiting time, manual-intervention frequency, and departure-target compliance.

Nikita Belenets Business Analyst @ Wildberries

“Artem helps the team tackle the most important challenges: large-scale changes, new product solutions, and urgent problem-solving. He quickly dives into complex business problems and delivers practical, well-thought-out solutions under tight deadlines.”