How to Improve Accuracy in Dragon NaturallySpeaking

The Dragon Newsletter has a tip on how you can improve accuracy and reduce word recognition errors in Dragon Naturally Speaking:

When Dragon occasionally misrecognizes a word, it’s tempting to simply use the keyboard to delete the text and type in the correct word or phrase. Similarly, the “Scratch that” command makes it fast and easy to delete the transcription of the previous utterance, or string of dictated text. However, overuse of this command may keep you from achieving even higher levels of accuracy with Dragon. This is because simply deleting the text doesn’t allow Dragon to “learn” from the misrecognition.

Instead of saying “scratch that” when you dictate text that is incorrectly transcribed by Dragon, a better choice is to select the misrecognized text (manually or by voice) and perform the correction (again, either manually or by voice) so that Dragon can learn from the error. Dragon will be less likely to misrecognize that word or phrase the next time around if you take the time to correct it. A direct command option to correct misrecognitions is “Correct XYZ,” where XYZ is the word or phrase you want to correct.

Amit Agarwal is a web geek, solo entrepreneur and loves making things on the Internet. Google recently awarded him the Google Developer Expert and Google Cloud Champion title for his work on Google Workspace and Google Apps Script.

Awards & Recognition

Google Developer Expert

Google Developer Expert

Google awarded us the Developer Expert title recogizing our work in Workspace

ProductHunt Golden Kitty

ProductHunt Golden Kitty

Our Gmail tool won the Lifehack of the Year award at ProductHunt Golden Kitty Awards

Microsoft MVP Alumni

Microsoft MVP Alumni

Microsoft awarded us the Most Valuable Professional title for 5 years in a row

Google Cloud Champion

Google Cloud Champion

Google awarded us the Champion Innovator award for technical expertise

Want to stay up to date?
Sign up for our email newsletter.

We will never send any spam emails. Promise 🫶🏻