Who could use this feature?
The Audio Transcription feature is designed for all who would like to automatically transcribe the entire conversations, word for word. Instead of manually writing detailed notes during or after the conversation, the system can automatically convert spoken dialogue into written text.
In many field situations, reporters must both conduct a conversation and document it at the same time. This can make it difficult to capture all relevant details or maintain a natural conversation with the person being interviewed. Audio transcription helps by automatically generating a written summary of the discussion.
AI speech-to-text technology converts recorded audio into text, allowing professionals to focus on the conversation rather than taking extensive notes. You can benefit from it by:
Capturing more complete and accurate information
Reduce the time needed to write reports after a visit
Create reliable documentation of conversations
Audio Transcription helps reporters spend more time on meaningful conversations and less time on writing reports, while ensuring that important information is properly documented.
How to use it?
To be able to use the audio transcription feature, administrators must use an "Audio" question type when setting up the questionnaire, so that everyone can start using it.
Then during an investigation or conversation, reporters can record the discussion with their smartphone or tablet by any voice recorder. Once the conversation is recorded, user can upload the audio file to Montr Rapportage App and the system automatically generates a written transcript of what was said.
After uploading user gets notified about the upload status and it takes a few minutes for the conversation transcript to be generated. Typically, the processing time - depending on the length of the recording - is 5 minutes per 20 minutes of audio recording. However, due to high volume, processing may occasionally take a little longer.
As soon as the transcript is ready, the user receives a notification via email. Below is an example of such a transcript in the Montr app:
The reporter can then review the transcript, make small adjustments if needed directly in the app - the transcription will naturally never be 100% perfect, and will undoubtedly contain some incorrect wordplay.
After that, it can be finalized and converted into a PDF document. This can then be easily shared with the other participants in the conversation.
Please note the following:
The quality of the audio recording determines the quality of the conversation transcript. So make sure the recording device is placed close to and between the speakers.
If people speak poor Dutch, this will inevitably affect the transcript. It is therefore always a good idea to review the transcribed conversation report afterward and make minor corrections where necessary.
It is also important to inform the conversation partner in advance that the conversation is being recorded and to ask if they have any objections.
Because the verbatim transcript is made available to both parties, this promotes transparency. In this way, there can never be any disagreement afterward about what was discussed.
The technology is based on the so-called Whisper AI speech-to-text model. The implementation, incidentally, has been designed with a strong focus on privacy. The input (audio) is not stored in the speech-to-text model, nor is the audio used to further train the model. And, of course, the transcription is performed within the EU.
Would you like to test and/or implement this feature? Let us know; we’d be happy to help you set it up in the questionnaire. Please contact us at helpdesk@montr.nl