Automating ticket quality control with Vestel and Commencis on AWS
Learn how Vestel and Commencis built an AI solution for quality control on AWS for around 8,000 service tickets a day.
Benefits
- of eligible tickets expected to be automated
- 80%
- weeks to complete full implementation
- 6
Overview
Vestel, a subsidiary of Zorlu Holding, reviews around 8,000 after-sales service tickets a day in Türkiye. As manual quality checks became harder to scale, the company used Amazon Web Services (AWS) to build a serverless, event-driven solution. Developed collaboratively by Vestel and AWS Partner Commencis, the solution is expected to automate up to 80 percent of eligible tickets and reduce manual-review workload.
About Vestel
Vestel is a Türkiye-based global technology company, delivering electronics and home appliances to over 160 countries.
Opportunity | Scaling quality control as service tickets multiply
Vestel has been one of Türkiye’s leading exporters in the electronics sector for 28 consecutive years. Its ticket records include product and service data, technician notes, parts and labor details, images, videos, and supporting documents. The evaluation criteria for each ticket vary by product group, service operation, and evaluation phase, requiring manual review to interpret both business rules and unstructured evidence.
As ticket volume increased, Vestel explored new ways to evolve its quality control processes. The company sought a solution that could scale while maintaining decision quality, explainability, and human oversight.
Solution | Combining business rules and AI for ticket evaluation
Building on an existing relationship, Zorlu Holding collaborated with Commencis for its AWS and AI expertise. After running a successful proof of concept, the parties jointly deployed the use case, completing the full implementation in 6 weeks.
They designed the solution as a serverless, event-driven system. For traceability and reprocessing, Vestel stores support data from customer relationship management in Amazon Simple Storage Service (Amazon S3), an object storage service. AWS Lambda functions provide serverless compute for ingestion and transformation. And Amazon DynamoDB—a serverless, fully managed, distributed NoSQL database—stores ticket states, evaluation results, and configurable criteria for both operational processing and dashboard access.
The architecture separates deterministic business rules, such as date checks and numerical thresholds, from evaluations that require semantic reasoning. Those semantic evaluations run on foundation models through Amazon Bedrock, a solution for building generative AI applications. To coordinate the evaluation workflow across text and images, the solution uses AWS Step Functions, a service for building and orchestrating distributed applications.
Outcome | Shifting reviewer focus to higher-value work
Establishing a scalable, measurable quality control workflow for around 8,000 tickets a day, Vestel expects the solution to significantly reduce manual-review effort while supporting a cost-efficient operating model.
“Built on AWS, the solution is designed to automate up to 80 percent of eligible tickets, routing complex or ambiguous cases to human reviewers,” says Ezgi Tosca, AI program manager at Zorlu Holding. “The project team successfully implemented the solution architecture and supported timely delivery.”
About AWS Partner Commencis
Commencis is a Turkish AWS Premier Tier Services Partner with over 450 experts in AI, cloud, design, and engineering.
Built on AWS, the solution is designed to automate up to 80% of eligible tickets, routing complex or ambiguous cases to human reviewers.
Ezgi Tosca
AI Program Manager, Zorlu HoldingAWS Services Used
Did you find what you were looking for today?
Let us know so we can improve the quality of the content on our pages