Summary

  • aiOla, an Israeli startup, has launched Jargonic, a speech-to-text model that can handle specialised jargon, background noise and diverse accents without the need for extensive retraining.
  • Jargonic has three key challenges in speech recognition – jargon, background noise and accents – and has been developed to handle these issues without the need for additional training.
  • The model achieves a recall rate of 89.3% on specialised financial terms, and is optimised for the noisy, real-life environments of the enterprise.
  • aiOla sees Jargonic as a step towards a broader shift in how humans interact with technology, and believes voice recognition will become the main interface for human-computer interaction in the future.
  • Jargonic is available via API to enterprise customers and integrates easily with workflow, applications and customer-facing services.

By Carl Franzen

Original Article