IBM and ElevenLabs add voice capabilities to watsonx Orchestrate
IBM and ElevenLabs are integrating speech tools into watsonx Orchestrate, IBM's agent orchestration platform. The update brings ElevenLabs text-to-speech and speech-to-text capabilities to enterprise customers building voice-based agents for customer service, internal support, and workflow automation.
The integration supports 70 languages and access to more than 10,000 voices. IBM began rolling out the capability in March, with regional availability across data centres including Frankfurt and London.
What this means for customer support teams
Voice agents built on the combined platform sound less scripted than earlier systems, making them more useful in real operating environments. The update targets customer-facing workflows including support and sales, as well as employee-facing tasks where multilingual access matters.
For AI for Customer Support teams, the addition means handling phone-based interactions, service-desk automation, and workflow tasks without building a separate voice infrastructure.
Enterprise controls take priority
IBM and ElevenLabs are positioning the offering around governance and compliance, not just speech quality. The platform includes PCI-aligned payment protections, a Zero Retention Mode for HIPAA-compliant handling, and data residency options.
This matters because voice systems in banking, insurance, healthcare, and government typically fail on governance and auditability rather than demo quality. Regional hosting and policy compliance are often as important as the speech model itself.
Part of IBM's broader strategy
This partnership fits IBM's approach of pulling specialist speech technology into watsonx Orchestrate rather than building every layer in-house. IBM previously integrated Deepgram's voice AI into the platform, signalling a pattern of partner-led development.
For ElevenLabs, the deal provides access to larger regulated deployments where procurement and compliance requirements limit which vendors organisations can use. For IBM, it makes watsonx Orchestrate more competitive in organisations that need phone and service-desk automation.
Mati Staniszewski, ElevenLabs co-founder, said voice is where AI agents either earn trust or lose it. Nick Holda from IBM framed the tie-up as part of the company's open ecosystem strategy, giving enterprise customers more choice in models and tools.
Speech-To-Text integration is now live across IBM's cloud regions, with phased availability continuing across additional locations.
Your membership also unlocks:
