Use case: Data generation for ticketing system using Synthetic Data Studio

This example illustrates how to generate a synthetic dataset for an agent ticketing use case.

Scenario Overview:

In this scenario, a user interacts with a customer support system by asking questions that require the system to direct the user to the appropriate resource. However, due to the absence of readily available training data or privacy restrictions on customer data, synthetic data is created to simulate both user questions and system responses.

Objective:

The synthetic data generation process enables knowledge distillation from larger proprietary datasets while addressing privacy concerns. This approach allows a smaller language model (SLM) to be fine-tuned to respond effectively, without relying on sensitive customer data.