Pexels (CC0), Pixabay

Technology adapts to needs. It is, in this way, much like the broader market. A customer needs something, or doesn’t know they need something but is using a few things to accomplish one task, and technology comes in to do what they need. It doesn’t stop here because this is just the start of the cycle. Once a technology is introduced, the customer uses it which then becomes a series of updates, new needs, and another iteration, and on and on and on. It is by no means a simple trajectory to track. Transcription services are indicative of this larger trend because there being a lot of different problems it can solve. There are different types that customers need to be aware of.

Human and AI

Firstly, there is the age-old comparison to be made between humans and computers. Let’s begin.

There are human transcriptionists and AI transcription software. Humans have been turning audio into text for millennia, centuries. It is an ancient skill. One human has done pretty well. However, as is natural, technology develops and can perform the task, in most cases, as well as a human and often more efficiently. There is plenty of optimism backing the idea that AI will not necessarily take the jobs of humans entirely but work alongside them to reach high levels of performance.

In the transcription industry, this rings mostly true. There are plenty of transcriptionists. Often, human transcriptionists are known for their accuracy and their reliability. However, they are the more expensive option. AI technology is still very accurate: Verbit’s transcription software has ninety-nine-percent accuracy. They can be used for live and pre-recorded content with similar reliability to humans but can continue to learn and potentially become more reliable as they become more informed.


There are different styles of transcription, also: Verbatim, Intelligent Verbatim, and Edited.

Verbatim is, as the name suggests, a transcription of exactly what is said, including pauses and filler words like “erm”, “uh”, and “hmm”. The idea is that it is possible. This type of style is often used in linguistic studies where academics will analyze a speaker.

Intelligent Verbatim is like Verbatim but with a few edits. There will be no filler words, but the transcriptionist will correct grammar and syntax and delete words that aren’t quite right.

Edited is the most stylized transcription. It is verbatim, but it will be heavily edited if needs be. The design is that it is of a publishable standard.

The decision of which to use often falls down to how the message wants to be communicated. More publishable work will be appropriate as press releases or social media posts as they will be reflective of a brand’s image.


Transcription services delivered by certain businesses will deliver interactive extras and out-of-box integrations. This will make things a lot easier once the live or pre-recorded video has finished. Often these transcriptions are used so that the audience or the speaker doesn’t need to make copious notes. They are able to sit and fully engage with what’s being said without having to worry about how they’ll remember it afterward. The convenience of having the text file drop into DropBox or Google Drive and be able to search for specific words and phrases to find specific parts of the audio for reference is invaluable.

The specific details of transcription are important as each case will have different needs. It’s like buying a car: a business won’t want a Tesla Model 3 to haul goods across the country. Needs must be known as much as what the service is capable of.