This paper introduces a unique implementation of Whisper by OpenAI, intended primarily for Telco applications. Whisper, initially released with support for 98 languages [6], represents a significant stepping stone in the world of Natural Language Processing. In this paper, we outline a hybrid model approach that enables a more diverse and cost-effective adaptation of the model for industrial deployment. We evaluate the C++ version against Whisper's original Python model, demonstrating comparable accuracy with a significantly reduced computational footprint. Our results show that the hybrid Whisper-Transformer model achieves effective real-time transcription while maintaining contextual accuracy, representing a scalable, cost-effective ASR solution for industrial applications.