Zendesk Embeds Real-Time Voice Translation Into Contact Center

Zendesk will offer real-time voice translation directly within its contact center platform, allowing agents and customers to speak and hear calls in different languages without switching channels or running separate applications.

The company announced Thursday that the capability will enter a closed early access program in October for eligible Contact Center customers, with general availability targeted for the first quarter of 2027. During early access, the service will support 13 languages including Chinese, English, French, German, Hindi, Indonesian, Italian, Japanese, Korean, Portuguese, Russian, Spanish, and Vietnamese.

Why it matters

Real-time translation has remained expensive and operationally complex for most contact centers, typically requiring separate vendor relationships, desktop overlays, or custom integration work. By productizing translation as a native feature with unified administration and call controls, Zendesk removes technical and procurement friction that has limited adoption beyond high-value verticals like healthcare and finance. Organizations can now route calls based on issue type and agent expertise rather than language availability, potentially reducing multilingual staffing costs while expanding language support.

How the native integration works

The translation capability operates within Zendesk's existing agent interface. Administrators can enable it by queue and choose whether call recordings preserve original audio, translated audio, or both. Agents can identify a customer's language through the contact flow or select it manually, then control translation without leaving the call screen.

"The goal is to enable our agents and our customers to speak in the language they know, no matter what language the other person is speaking and hear the language they know as well," said Lilith Shoemaker, product marketing manager at Zendesk, according to No Jitter, which first reported the details.

The integration extends to AI agent handoffs. When a customer first interacts with a Mandarin-speaking voice AI agent, for example, an English-speaking live agent can receive an English summary with the language attribute automatically activating translation.

Differentiation from overlay models

Point solution providers including Krisp and Sanas have offered real-time AI voice translation for contact centers since 2024, typically through partnerships that require agents to run separate desktop applications. Sanas also provides a Chrome extension for browser-based platforms, though Market.us Scoop estimates only about 20 percent of contact center agent desktops worldwide use Chrome thin-client technology.

Both vendors offer developer SDKs for native embedding, but that approach generally requires custom integration work by enterprise IT teams or vendor professional services. Companies using these solutions must also establish direct commercial relationships with translation providers, even when installing through a CCaaS marketplace.

"The difference is that we've done this in our product as one piece. We've got all the context," said Nuri Gocay, director of Contact Center Platform Architecture at Zendesk. That context includes contact attributes and call events integrated into routing, call controls, and recording workflows.

Early access limitations and pricing

The early access version will not support multiparty or transferred calls, supervisor monitoring, barge-in functionality, or video calls. It also lacks viewable usage reporting and administrative usage limits. Customers will not be charged during the program, though usage will be capped.

Zendesk has not finalized pricing for general availability but confirmed it will use a usage-based model. The company plans to analyze early access program data before setting commercial terms.

No Jitter first reported the announcement and provided details on the early access program.