Empathetic Multilingual Voice Agents Powered by Generative AI in Healthcare: Development and Implementation of CoSpeak and Empathia

Background: The integration of generative AI-powered voice agents in healthcare is revolutionizing patientpractitioner interactions, enhancing clinical workflows, and improving overall patient care. Methods: We developed and implemented two intricate generative AI models, CoSpeak and Empathia, supporting multilingual voice-to-voice communication in English, Spanish, French, Hindi, and Telugu. These agents decode human emotions and exhibit empathy, providing adaptive and personalized interactions. Results: Experimental results from pilot projects demonstrated significant improvements in patient satisfaction (90%) and operational efficiency (30% reduction in intake time). Emotion recognition accuracy reached 88%, with ASR achieving 95% accuracy across supported languages. Conclusions: The implementation of empathetic multilingual voice agents shows promising results in improving healthcare communication and patient experience