Speech to Text Transcription

Cabertel's Speech to Text transcription feature is used to transcribe your voicemails into textual form which is then included in the email we send you whenever you receive a new voicemail. The software we use for transcribing voicemails utilizes Artificial Intelligence (AI) to learn over time, thus continuously improving the quality of the transcriptions.

Transcription Benefits

Here are some of the benefits of using transcription:

• As part of our voicemail to email feature you can receive voicemail audio attachments in your email. Now, by using transcription, you no longer need to listen to your voicemails, just read them instead. This is prefect when you are in an environment, such as a meeting, where you cannot listen to new voicemails.
• The speech to text feature utilizes complex software algorithms from industry leaders and is high quality. Transcription will likely never be 100% accurate, especially for heavy accents, but even 90% accuracy is enough to quickly determine the context of a voicemail.
• AI is used so the speech to text algorithms learn over time thus improving the transcription accuracy.
• Additional languages can be configured for speech to text transcription.
• The algorithms work on WAV, MP3 and OPUS audio files. In practice we typically compress voicemails into MPEG3 format and transcribe. This ensures that the voicemail to email audio attachment is small. The text is displayed in the message body of the email. The transcription is also displayed in your online Web portal.
• Phone numbers and dates are automatically converted from text to numbers, as can be seen in the example below.

Transcription Example

An example of a transcribed message can be seen below. You can listen to the voicemail by clicking on the audio link and below it you can see the text that was included in the email we received. In this example it is 100% accurate and note that the date and the Cabertel phone number were automatically converted from words to numbers.

TEST: Voicemail Transcription Audio

TEST: Voicemail Transcription Text

Transcription Downsides

The only downsides to this technology is that it will likely never be 100% accurate and there is a delay in receiving your voicemail due to the time it takes our servers to process the speech to text transcription. In reality, this delay is usually less than 30 seconds. There is also a small cost associated with this feature as we have to license complex software algorithms.

Next Steps

If speech to text transcription is of interest to your business ask your dedicated account manager for pricing or use the form below to receive more information if you are not an existing Cabertel customer.