IBM and Deepgram are joining forces to integrate advanced speech-to-text and text-to-speech capabilities into watsonx Orchestrate. The collaboration aims to help organizations further automate their operations and better respond to the growing demand for conversational AI technology.
IBM is making Deepgram its first speech partner. The integration brings speech AI to the watsonx Orchestrate platform, enabling companies to build digital agents that respond to natural speech. Users will have access to advanced speech recognition, real-time captioning, and customization options.
Challenges in real-world audio
IBM sees organizations that implement AI-powered speech-to-text systems regularly encountering challenges. Background noise, diverse accents, and natural conversation dynamics make automatic transcription complex. Deepgram’s technology aims to overcome these obstacles by supporting dozens of Arabic and Indian variants, as well as regional accents.
Watsonx Orchestrate offers companies ready-to-use AI agents that can be implemented in minutes rather than months. The platform supports multi-agent orchestration and includes a catalog of pre-built agents and tools. The integration with Deepgram now adds speech functionality.
Applications in healthcare and financial services
The combined technologies open up opportunities for improved automated customer service, call analysis, and voice-driven data entry. The solution offers new opportunities, particularly in sectors such as healthcare and financial services, where accuracy and compliance are crucial.
By embedding Deepgram in watsonx Orchestrate Agent Builder, IBM customers can build voice agents and voice-driven workflows on a real-time foundation. Nick Holda, Vice President of AI Technology Partnerships at IBM, emphasizes that the integration introduces new possibilities for speech recognition and transcription. “This collaboration aims to help enterprise organizations accelerate their AI initiatives and reinforces IBM’s open ecosystem,” says Holda.