AI-Based IVR
Gassyrbek Kosherbay, Nurgissa Apbaz
TL;DR
The paper tackles IVR inefficiency by integrating AI components for Kazakh-language call handling. It proposes an end-to-end pipeline using Whisper-based ASR with LoRA fine-tuning on 558 hours of Kazakh data, IrbisGPT-based classification with RAG, and mms-tts-kaz for voice feedback, deployed via Ollama on local hardware. The practical results show significant improvement in recognition accuracy and routing efficiency, with $WER=16\%$ on Kazakh test data after adaptation, compared to higher baselines. The approach is adaptable to other languages and can be extended to further enhance interaction quality in real-world call centers.
Abstract
The use of traditional IVR (Interactive Voice Response) methods often proves insufficient to meet customer needs. This article examines the application of artificial intelligence (AI) technologies to enhance the efficiency of IVR systems in call centers. A proposed approach is based on the integration of speech-to-text conversion solutions, text query classification using large language models (LLM), and speech synthesis. Special attention is given to adapting these technologies to work with the Kazakh language, including fine-tuning models on specialized datasets. The practical aspects of implementing the developed system in a real call center for query classification are described. The research results demonstrate that the application of AI technologies in call center IVR systems reduces operator workload, improves customer service quality, and increases the efficiency of query processing. The proposed approach can be adapted for use in call centers operating with various languages.