Predict earlier
Flag hard-to-dispatch requests before assignment.
HomeMenu
Karnameh Forced Dispatch
Exploring why some vehicle inspection requests were less likely to be accepted.
Urgency was not the whole story.
13.54%Urgent requests forced
15.14%Non-urgent requests forced
31.8%Non-urgent with <1h remaining
01Predict dispatch risk
02Offer targeted incentive
03Assign before escalation
Vehicle type, timing, location, earning potential and capacity compounded. A targeted intervention gave operations an earlier decision point.
THE MODEL
Difficulty is contextual.
Force Dispatch appeared when several pressures stacked—not because of one request label.
01Vehicle
02Time
03Location
04Economics
05Capacity
RISK BY TIME OF DAY
The risk rose as the day progressed, peaking between 16:00 and 18:00.
Improve attractiveness
Use targeted incentives where economics are weak.
Shape demand
Offer alternative time slots before capacity tightens.
Match smarter
Use vehicle, area and capacity signals together.
VALIDATED IMPACT
+12%Acceptance with targeted incentives
23%Accepted an alternative slot
71%High-risk cases identified
68%Prediction precision
The service
Customers book an on-site vehicle inspection for a vehicle, location and time. The request then moves through dispatch, acceptance by the field network and the inspection itself.
The question
Some requests were not accepted through the normal assignment flow and required extra operational work. What makes a request harder to dispatch?
What the data revealed
Urgency alone did not explain the pattern: the Force Dispatch rate was 13.54% for urgent requests and 15.14% for non-urgent requests. Nearly 31.8% of non-urgent requests had less than an hour left.
Reframing the problem
Vehicle complexity, time of day, location, expected earnings and capacity could stack. The resulting question became: how might we predict and reduce dispatch difficulty before assignment?
Designing the intervention
The team tested a risk signal before assignment and targeted incentives for the hardest requests. The model identified 71% of high-risk cases at 68% precision, helping operations intervene earlier instead of escalating after a request stalled.
Outcome
Targeted incentives lifted acceptance by 12% and cut Force Dispatch by 18% in the intervention. Across the initiative, the study reports 16% fewer Force Dispatch cases, 9% higher acceptance, 18% faster assignment and 14% fewer manual interventions.
−16%Force Dispatch overall
+9%Acceptance rate
−18%Assignment time
13.54%Urgent requests
15.14%Non-urgent requests
31.8%Non-urgent with under 1 hour left
REAL PROJECT SCREENS
From the original project.
INTERACTIVE PROTOTYPEOpen prototype in Figma ↗Opens in a new tab