As 2023 fast approaches, the resilient language services industry finds itself pulled in several directions; not the least of which is a path carved by rapidly changing technology. In fact, the language services marketplace is beginning to resemble a create-your-own-sandwich menu, with adjacent services increasingly becoming a part of it.
We asked readers whether language service providers (LSPs) have to diversify into non-translation areas — audio-visual services, cultural and linguistic consulting, testing, and multilingual content origination, among others — in order to thrive long term. Most (41.4%) agree that is the case.
To a not-so-distant percentage of readers (31.0%), adjacent services are an option for LSPs, but not a must. Service diversification is a matter of choice in strategies for less than a quarter of respondents (15.5%), and the rest (12.1%) believe remaining focused [on a service] is a viable long-term strategy.
DeepL Still Beats Its Own MT Drum
DeepL has held our collective attention for many months now, and justifiably so. The fastest-growing machine translation company of 2022 has been taking on additional investment putting its valuation at about USD 1 billion.
We asked readers if they thought five-year-old DeepL still retains its original quality edge for non-custom MT. A vast majority (56.9%) thought it did. A little over a quarter of poll responders are not quite as sure (27.6%). For the rest (15.5%), DeepL’s quality edge is past.
Prepped Source for Better Transcreation?
Specifically, Savenkov spoke of making source content more general by removing cultural content (“neutralizing”) from English copy before transcreation. Internationalization, you say? Yes, only Intento has experimented doing this with a number of tools across the production cycle.
We asked readers if neutralizing marketing source content does indeed facilitate a more creative translation. Respondents were almost evenly split in their views of the matter, with a third saying it is conditional on the type of text (36.0%). A third of readers voted Yes (32.0%), while the rest held the opposing view (32.0%).
Adopting Automatic Transcription
Scott Stephenson, CEO of Deepgram, joined SlatorPod in November 2022 to discuss the many ways in which his company is using speech recognition and related technologies to greatly accelerate real-time multilingual transcription and add to it a “next level of understanding.”
Beyond the ways in which Deepgram is advancing automatic transcription with a deep learning speech API, the technology seems to have become one of those things that most people know exist, but is only used by those with a prescribed need.
Our poll results certainly reflect a lack of adoption of automatic transcription. We asked readers if they use the technology in internal meetings or client calls, and the majority (59.3%) answered they never do.
Close to a third (28.1%) of responders said they rarely use automatic transcription; two small groups were evenly split between using it sometimes (6.3%) and always using it (6.3%). No one said they use it often.