How to Correct Automated Transcription in Microsoft Teams and Zoom

A very common quote request at the moment into TP Transcription and University Transcriptions is for the correction of the automatic (AI) transcription provided by Microsoft Teams and Zoom. These services have been offered by Microsoft and Zoom for some time now, but the quality of the subtitling or transcription remains highly inaccurate, particularly when the meetings/recordings involve strong accents.

We have had a number of requests in the last few weeks from researchers looking to get their automated transcriptions corrected, and asking us for assistance and advice in the best way of doing it.

Use the Automated Transcription as a Starting Point?

In an ideal world it would be simple enough for the clients to simply listen to the recordings and make the occasional corrections, but what we have found with the automation is that the transcriptions are so bad it’s often cheaper and easier in terms of time and expense for the client to send us the full Zoom or Teams file and we transcribe the whole thing again from scratch. If the recording is for more than one person we have found that the automatic transcription is so appalling that there is no merit at all in our transcribers working from it. It is easier for them to simply start again from scratch.

The Quality of Automated Transcription – Not Yet Usable

This is a recurring theme over the past few years with Silicon Valley inventing automated or AI transcription, but for the AI transcription to not be up to scratch for anything other than dictation work. Voice recognition and automated transcription has been a service provided by the likes of Dragon since the early 1990s and there has been nothing new onto the market ever since in terms of being able to accurately transcribe multiple speakers and/or multiple speakers with an accent.

Starting From Scratch

If you have a Microsoft Teams or Zoom audio recording that you need a transcription from or subtitling, the best way of getting this completed is to pay for a full manual transcription, unless the recordings are of one person dictating or speaking slowly and clearly directly into their PC or phone. Anything else will result in a mass of indecipherable incoherent sentences that if you try to correct will take more time and effort than it would simply to send off the recording to get it transcribed in full. Our transcribers have done so many of these now that we are able to confirm that in just about every case the automatic transcription has been a complete waste of time.

High Quality Manual Transcription Remains the Best

We specialise in hard to hear recordings as well as strong regional accents and we do not use automated transcription at all as part of our service. Our transcribers are all UK based native English speaking transcriptionists with many years experience behind them. We also proof-read and quality check work.

So in summary the easiest way of getting an automated transcription from Zoom or Teams corrected is to send the Zoom or Teams file directly to TP Transcription Limited using our upload file service, and we will transcribe the recording for you. Don’t bother with the automatic transcription option unless the audio is a recording of one person only. For meetings of two or more people it is usually a complete waste of time.

If you do not know how to access your recording in Teams or Zoom, please see our blog articles here – and here –

Our Accreditations

We are Cyber Essentials Plus audited annually and we hold the Cyber Essentials and Cyber Essentials Plus certificates. We are UKAS ISO 27001:2022 audited and accredited and ISO 9001 & ISO 14001 systems accredited company. We are members of the American Translators Association and we are assessed for GDPR compliance annually by IASME (Cyber Assurance Level 1).

10% Profits to Charity

10% of our profits are donated to the Ten Percent Foundation, a charitable trust registered in the UK. Since 2000 over £150,000 has been donated to projects in Africa and the UK. Click here for details.