Case Study: Re:Drink achieves 100% dispenser uptime with Nixtla's TimeGPT

How Re:Drink improved refill forecasting and maintained 100% dispenser uptime with Nixtla's TimeGPT, saving 1 hour per day on manual forecasting

Re:Drink, a Munich-based startup that provides bottleless smart beverage dispensers, faced the challenge of needing to accurately predict when its dispensers required refill shipments. Manual forecasting was time-consuming, error-prone, and could not scale with their growing customer base, creating a risk of stockouts that would interrupt service. They turned to the vendor Nixtla and implemented its TimeGPT service to solve this.

By integrating Nixtla's TimeGPT via an API, Re:Drink automated its refill forecasting. The solution provided rapid, accurate predictions without the need for manual model management. This resulted in the team saving 30 hours per month on forecasting tasks, ensured 100% dispenser uptime by preventing stockouts, and created a scalable system to support their growing operations.

Re:Drink

Alex Colucci

Co-founder & Tech Lead