What is the quality of the instant translation

Artificial intelligence has made giant steps forward in the last 4 years, featuring GAFA (Google Microsoft Facebook) and a Deepl outsider, the progress is spectacular compared to 5 years ago.
Nevertheless, can we say that machine translation is completely mature in 2020?
The answer is a matter of debate, as it depends very much on the context of your business, quality requirements and the budget allocated.
Localizing software and technical documentation with the traditional approach is indeed expensive in the long term.

See a concrete case study.

Several approaches currently exist to try to reduce costs
- Human translators assisted by an automatic translator which reduces the cost per word
- A collaborative translation tool that can involve users and translators.

Human translation remains essential for content involving emotion or humour or being very technically advanced (legal for example)

On the other hand, machine translation may well be suitable for software and documentation whose content remains essentially confirmation messages, descriptive or informative text with little ambiguity.
In the latter case, the gain is truly revolutionary with a reduction of 200 to 300 of your costs.