Case Study: Zendesk achieves faster, more accurate customer service with Amazon Web Services (AWS)

Zendesk Enables Faster Customer Service Using Deep Learning on AWS

Zendesk, a SaaS customer‑service platform, needed to help customers find fast, relevant answers on their own and scale a self‑service support model without overburdening agents. The challenge was to build and iterate deep‑learning models quickly and handle large training datasets and compute requirements so answers could be delivered in seconds.

Zendesk built Answer Bot using TensorFlow on AWS—leveraging EC2 P2 GPU instances, Amazon S3, Aurora, and Amazon SageMaker—to accelerate training, deployment, and experimentation. The solution shortened development time, enabled on‑demand scaling for research and production, and delivers targeted answers that resolve many support requests before they reach agents, improving response speed and customer satisfaction for customers such as Dollar Shave Club.

Zendesk

Arwen Griffioen

Data Scientist