GetTransfer cuts costs with in‑house AI dispatch

The problem

GetTransfer is a ride marketplace balancing customer ETAs, driver utilization, and margins. As volume grew, manual and rules‑based dispatch struggled with peak demand, inaccurate ETAs, and poor matching—classic problems that hit SMB transport operators’ profitability and CSAT.

The AI solution

They built an in‑house AI dispatch that predicts demand, ranks drivers, and prices intelligently. The system ingests historical trips, geospatial supply, live traffic, seasonality, and cancellation patterns. Models produce driver match scores and ETA confidence, while business rules enforce safety and SLAs. Rollout started in one market, with A/B tests guiding feature toggles.

Results & metrics

• Lower cost per completed ride by improving first‑match success
• Higher on‑time pickup rates via more reliable ETA predictions
• Fewer cancellations and reassignments during peak windows

Lessons for SMBs

You don’t need a research lab. Start with the data you already have (orders, routes, demand patterns). Ship a small pilot, track on‑time rates, first‑match %, and cost per job, then iterate. Use a pragmatic mix of heuristics and ML.

Learn more

Read the original coverage for deeper context and numbers.

Looking for help building pragmatic AI into your ops? See our services.