Microsoft Previews AI Speech Translation App

On November 21, 2024, Microsoft announced the preview of Microsoft Translator Pro, an enterprise-level, standalone speech-to-speech (S2S) and document machine translation (MT) app available first to a few “applicants” that meet certain criteria, including having an Azure enterprise account.

The mobile translation app will be available initially on iOS 15 or later. Those who are granted access will be able to test its real-time speech translation and transcription features, along with offline language support (initially English, French, German, Italian, Japanese, Mandarin Chinese, Russian, Spanish, and Ukrainian).

The app includes an update to Azure’s AI Translator’s Document Translation (DT) feature. Available originally for digital document and scanned PDF translation, the functionality will now also support “mixed-content documents, translating both digital text and text embedded within images,” according to the announcement. This feature uses Microsoft Azure AI Vision API.

The announcement includes a video in which a person tests the speech-to-speech functionality on an iPad, using English and Mandarin Chinese in a hospitality-type simulated dialogue. A voice is heard translating English speech into Mandarin speech, but not Mandarin speech into English speech. The live text transcription and translation for both languages are shown on the right side of the video.

Other features highlighted in the announcement include granular control for IT admins over app deployment, conversation history, and data security. On the latter, the app is said to provide enhanced privacy for conversational data within Azure.

Beyond the mobile iOS device support, at face value, this Pro version of the Microsoft translator seems to differ on a few points from the business version of its current translator offer. One is the level of data privacy and confidentiality offered through an Azure tenant, which might be a key differentiator for business use cases.

Beyond the specific mention of data security and support only for iOS, the features listed in the Microsoft Translator Pro announcement are comparable to other standalone mobile translation apps like Samsung’s Live Translate and Google Translate.

To qualify for the preview, interested individuals must work for a company where English is the primary spoken language, have a valid Azure enterprise subscription with data storage in the US, meet the seven eligibility criteria, and agree to the terms of use. 

The app will be available in January 2025, but there is no clear indication of when access will be open to other Azure users.