Speechmatcis Adds Language Identification to Speech Recognition AI
Speech recognition technology startup Speechmatics augmented its speech-to-text platform with the ability to identify the language spoken in a media file. Speechmatics is offering the new Language ID tool as a way to more accurately transcribe files without the need for human review to sort which linguistic transcription model to employ.
Speechmatics primarily serves businesses with its speech recognition platform, promising highly accurate transcripts even in informal settings and with wide-ranging accents and speech patterns across millions of hours of conversation. That flexibility extends to more than three dozen languages. Matching the software to the languages spoken in a media file still required human selection until now, however. The Language ID feature skips that step. The AI can process and match the language spoken to its options, even if the human user isn’t sure what language is being spoken. A dozen of Speechmatics’ languages are identifiable by the new feature so far, with the others likely to follow.
“Up until now, identifying languages without human intervention has been costly and time-consuming for users of speech-to-text,” Speechmatics CEO Katy Wigdahl said. “However, with our new Language ID, this will be a thing of the past and allow customers to swiftly identify and transcribe media files – with less hassle and more efficiency. We can’t wait for our customers to use this Language ID and see it deliver accurate and valuable results.”
Demand for AI-powered transcription of meetings and media has accelerated the growth and feature development of companies like Speechmatics, which raised $62 million in June. Speechmatics counts many government agencies and international brands among its clientele, including Vonage, Veritone, and Deloitte UK. The technology is embedded in the products of its partners like navigation software platform what3words and call center platform Puzzel. Media brands with large audio archives that have not been filed unreviewed for a long time can use the new tool for transcription without requiring human aid in listening and marking different languages. Meanwhile, contact centers can better analyze the data of a multilingual customer base or individual customers switching languages.