Our client amassed hundreds of hours of audio data that needed to be processed, transcribed and categorized so that it would be usable for a voice recognition software that is featured on mobile devices and desktops.
Seamless Project Management Across Borders
iSoftStone recruited and trained teams in over seven countries simultaneously. Recruiting, coordinating, and training these resources in different time zones and languages within the limited timeframe and a high volume of files requiring transcription was a challenge that iSoftStone gladly accepted. iSoftStone was able to successfully complete transcription in all seven languages, ramping up to over 500,000 transcribed texts per week over a six month time period. iSoftStone’s work also included categorization tags, validation and QC that met the client’s required SLA’s and delivery standards.
Transformed from Ineffectual Audio to Tangible Insightful Data
iSoftStone quickly turned around millions of transcribed and tagged utterances from hundreds of hours of audio data. The client’s audio files were transformed into meaningful data that is now used to improve speech recognition technology in several key language markets. Not only did iSoftStone deliver on time, but all transcriptions were completed with a KPI accuracy rate of 98% for grammar, syntax, and para-linguistic tag use as well as 98% validation accuracy in seven language markets. Furthermore, iSoftStone was able to quickly ramp up a team to full production in less than two weeks with an adjustable schedule. iSoftStone’s speed and accuracy rivals any competitor in the rapidly growing 30 billion dollar voice recognition industry that has increased market share by 50% in the past year.