Using Natural Language Processing (NLP) and machine learning, Verbit’s proprietary speech recognition engine aims to comprehend, organize and generate speech-to-text captioning. Co-Founder and CEO Tom Livne says his company's solution is over 90 percent accurate. The entire process takes a matter of seconds according to Livne.
Such technology might be valuable for contact center managers hoping to improve CX.
Verbit's automated speech recognition engine is also self-learning and uses adaptive algorithms while feeding information from external sources and previous sessions back into the model. This ensures the engine is continually training for heightened accuracy.
Could Verbit help improve CX for contact centers?
The Israeli startup's mission seems to be - revolutionize real-time transcription and captioning to the highest possible degree.
In order to increase accuracy, (up to 99 percent) - the SaaS offering incorporates human transcribers to correct any mistakes. The solution then takes note of corrections made and feeds them into the engine. According to Livne, Verbit has a 10x faster turnaround time through ASR (automated speech recognition) technology and enables users to download transcripts just minutes after a session wraps up.
The process of transcription is simple: Upload audio/video files to Verbit Cloud, the engine does its job, human intervention occurs, then the file is then ready for download. This alone could make content for contact center training more accessible, engaging, and easy to search.
“Verbit's adaptive Artificial Intelligence technology automatically reduces background noise, identifies specific terms, and differentiates between speakers for the highest accuracy.
Some interactive features include searching inside a transcript; identifying, extracting and tagging keywords; connecting to related content; LMS and video hosting platform integrations; and the ability to not pre-schedule a reporter,” Livne said.
[caption id="attachment_18675" align="alignright" width="300"]
Tom Livne[/caption]