Voice Recognition Applied to Transactions

Processing power, improved networks  and integrations  have enabled advances in voice recognition.  Voice recognition has been adopted by call centers.  Interactive Voice Response (IVR) is now common place and with advances more accepted by callers.  Voice command and control is being implemented for home use (Apple HomePod, Amazon Echo, Google Home) with strong consumer adoption.   We expect to use voice input for commanding and controlling applications at home and in automobiles.  Voice to text systems shown in visual voicemail have improved substantially in recent years.

However to date voice has not been a primary medium for the input of  transactional data.   However there are some transactional applications that are being successfully employed – the most visible of these is transcribing medical records, which Nuance and M*Modal have invested heavily in.  The input is semi-structured and the application conceptually very similar to traditional paper based forms processing.  The physician is trained to follow a specific structure in identifying and diagnosing the most likely problem and the proposed treatment.  Rules are built to understand to a sufficient degree to substantially reduce transcription costs.

But other transcription services have been slower to adopt automation. For example legal transcripts remain mostly un-automated and bank traders’ voice recordings are mostly manually searched.

We believe the state of voice recognition is on the cusp for broader adoption and growth.  There will be many opportunities for pattern based transactional capture from unstructured voice records, not only to displace current transcription services but also to be adopted more broadly as voice becomes a more effective business process input medium.  HSA is the leader in identifying and understanding these opportunities.  For more information please email Harvey Spencer Associates.

Leave a Reply

Your email address will not be published. Required fields are marked *