The system significantly reduced manual effort, enabling the client to discover and prioritize new customer needs with greater speed and accuracy, reaching 107 intents.
Meet our client
Client:
Industry:
Market:
Technology:
Client’s Challenge
A leading telecommunications company needed to improve their online support chat, which relied on an intent classification model trained within the IBM Watson platform. The model could detect around 90 manually identified intents, but the client sought to automatically discover new, previously unsupported intents.
Our Solution
We developed a GenAI-powered suite of tools designed to automatically analyze customer utterances and identify new intents. Using OpenAI GPT-4 Turbo and embeddings, we implemented the solution within Google Cloud Platform’s Vertex AI Workbench, making it accessible even to non-technical users.
Client’s Benefits
The system significantly reduced manual effort, enabling the client to discover and prioritize new customer needs with greater speed and accuracy, reaching 107 intents. Early results indicate that the client is now better equipped to stay ahead in a competitive market, with faster model updates.
