IBM partners with Deepgram to integrate Deepgram’s speech technology into IBM’s watsonx Orchestrate platform. Through this collaboration, Deepgram will become IBM’s first speech technology partner.
Under the agreement, Deepgram’s speech recognition and conversion technology will be embedded into IBM’s platform for building and managing AI agents and automation workflows. This move aims to lay the foundation for enterprise-wide transcription, real-time subtitles, and natural voice interactions for digital agents.
Both parties stated that this partnership will support the development of real-time voice-enabled agents and workflows in areas such as customer support, call analysis, and voice data input. It reflects the enterprise’s desire to simplify interactions with software agents. According to Grand View Research Inc., the voice and speech recognition market is expected to grow from $20.2 billion in 2023 (approximately 29.568 trillion Korean won) to 2030, with an average annual growth rate of 15%.
Deepgram’s technology focuses on solving challenges like noise, multiple accents, and real-world conversations. It supports 35 languages, including various Arabic and Indian dialects. Deepgram can deliver over 90% accuracy with latency below 300 milliseconds.
This integration aims to expand watsonx’s capabilities. Watsonx includes tools for building, training, and managing AI models, as well as orchestrating AI agents across business processes. IBM positions watsonx as its core platform for delivering open, enterprise-grade AI, integrated with hybrid cloud environments.