Viston helped us cut fuel costs and improve on-time delivery by implementing demand forecasting and AI-powered route optimization.

Business challenge

On-time delivery at 82%, volatile demand planning, and rising fuel costs with static, dispatcher-driven routes.

Our approach and solution

Viston built SKU/DC-level forecasts and a VRP optimizer (time windows, capacity, driver constraints) with live telemetry and traffic-aware re-optimization.

Driver app integrations provided ETAs, exception capture, and geofenced proof-of-delivery.

AI applications and benefits delivered

Time-series forecasting, reinforcement-informed routing heuristics, and ETA prediction.

Benefits: fewer empty miles, better OTIF, and lower fuel/CO2.

Cost of implementation

USD 220,000 including data engineering, model ops, and mobile integration.

Time to implement

12 weeks to first 8 depots; national rollout in 6 more weeks.

Tools and technologies used

Databricks + Spark, PyTorch Temporal Fusion Transformer, Facebook Prophet for baselines, Google OR-Tools, Google Maps Platform, Airflow, Postgres, Superset, MQTT telemetry.

Quantitative outcomes

Fuel cost per km reduced 17%.

On-time delivery improved to 95%.

Empty miles down 14%.

CO2 emissions reduced 12%.

2.4x ROI in 12 months.

Key performance indicators (KPIs) tracked

OTD/OTIF, fuel per km, km/stop, MAPE for demand forecast, route adherence, average delay minutes.

Pre- and post-implementation metrics

OTD: 82% → 95%.

Fuel/km: -17%.

Empty miles: -14%.

Forecast MAPE: 28% → 14%.

Stakeholder quotes or testimonials

“Dispatch stopped being guesswork. The optimizer reacts to reality, not yesterday’s plan.” — COO

Regulatory or compliance considerations

Brazil LGPD, telematics data minimization, retention policies, ISO 14001 tracking for emissions.