Local glossary
H
Huw Thomas
The product keeps getting proper nouns wrong. Names of colleagues, local hospitals and local place names are invariably incorrectly transcribed meaning they need to be adjusted. It would presumably be possible to have a cache of frequently used names that each user could update to reflect local needs.
H
Huw Thomas
The most common mistakes are colleagues with unusual surnames, followed closely by local place names. If we had a customisable list of local names that the algorithm could refer to we could update them manually. Does the AI know where we are? If it does, it should be able to deduce local placenames and hospitals etc.
Tom
Hey Huw Thomas, thanks for your feedback! I have a few more questions for you:
- What specific names or types of names are most frequently transcribed incorrectly?
- How would you prefer to update or manage this cache of frequently used names?
- Are there any specific local names or terms that should be prioritized in the transcription process?