Does Speech Anywhere learn my voice?

Overview

This article explains how Speech Anywhere improves recognition over time.

Applies to

All Users

AI-powered voice learning

Speech Anywhere is powered by Microsoft Azure Cognitive Speech Services, which uses deep learning to adapt to individual users over time. The engine improves its understanding of your accent, speech patterns, pace, and vocabulary the more you use it.

Speech to text replacements

In addition to the underlying AI learning, you can actively improve accuracy by adding Speech to Text Replacements for specialist terms, proper nouns, abbreviations, and frequently misrecognised words. These replacements are stored against your account and applied every time you dictate. See the knowledge base article “How to add a Speech to Text Replacement?” for guidance.

Consistent correction

When you correct errors in transcribed text, this feedback contributes to improving future recognition. It is important to actively correct mistakes rather than leaving them, as this helps the system refine its model for your voice.