Cartesia's 2024 State of Voice report details significant advancements in voice AI technology, highlighting breakthroughs in conversational AI systems that combine speech-to-text (STT), large language models (LLM), and text-to-speech (TTS) to facilitate natural, real-time interactions. The report discusses the emergence of new model architectures like Sonic TTS, which enhance deployment flexibility and efficiency, and highlights the evolution of voice AI APIs that replace traditional systems with dynamic, enterprise-scale solutions. Voice agents have expanded across various industries, from loan servicing and healthcare to logistics and hospitality, streamlining business functions and supporting more complex tasks with improved reliability. The report anticipates the growing role of compact, on-device models in enabling local processing and privacy, as well as advances in fine-grained control of synthetic speech, which will further integrate voice AI into diverse workflows and entertainment experiences. As the industry progresses, 2025 is expected to see more sophisticated and accessible voice AI systems, driven by innovations in neural network architectures and enhanced model performance.