Microsoft is adding automated transcription to Word on the web, giving Microsoft 365 subscribers unlimited transcriptions of recordings made through the program at no extra cost, and up to five hours of free transcription per month for audio recorded and uploaded separately.
The new feature, announced this morning, will compete with free and paid automatic transcription services and programs such as Otter.ai, Temi, Google Live Transcribe, Nuance’s Dragon software, and others. There are hundreds of millions of Microsoft 365 subscribers who have access to Word and other Office programs, increasing the potential impact of Microsoft’s move.
Advances in artificial intelligence and speech recognition have made automatic transcription a viable, if not yet fully accurate, alternative to human transcription in many scenarios. The shift to virtual meetings has also fueled the demand for automatic transcription, as users take advantage of built-in recording features in video conferencing services.
In another high-profile example, Zoom and Otter.ai rolled out a transcription feature as part of paid subscription plans earlier this year.
As a feature of Word on the web, Microsoft’s new automatic transcription will work on any computer and with any meeting software, including but not limited to Microsoft Teams, the company said. Automatic transcription is slated to be added later this year to Microsoft Word apps for iOS and Android. The feature could come to Microsoft Word for desktop in the future, as well.
Microsoft’s transcription feature is available starting today to Microsoft 365 subscribers under the “dictate” icon in Word on the web. When active, the feature appears in a pane in the right-hand side of the browser window, offering the ability to start recording a recording for transcription through the program or upload recordings made separately.
The transcript appears first in that right-hand pane, with the option to edit the text and copy the full transcript or individual paragraphs into the main Word document.
The company decided not to display the transcription in real-time for now, opting instead to show the transcribed text after the recording is completed. That decision was the result of user testing, although Microsoft is open to considering real-time display of the transcription in the future, said Dan Parish, Microsoft principal group PM manager, in a briefing with reporters this week.
The transcription is taking place in the background, reducing the time required for processing after the conversation is completed to a matter of minutes, the company says.
When audio is being recorded, there is a red icon in the browser tab, and a timer indicating that the recording is active, but one downside we noticed in our initial usage is that there is no sound meter or other means to confirm that the audio is actually registering on the recording, which could be a concern given the fickle nature of microphone settings in many computers.
In addition, there is currently no option to purchase additional minutes for uploading beyond the limit of 300 minutes per month, but the company says it is considering the option to pay for extra upload minutes in the future. There is no limit on the transcription of audio recorded directly through Microsoft Word on the web.