I am at the TAUS user conference in Seattle this week. Many companies are here to showcase their companies and technology. First up in this afternoon’s session is Speaklike. Launched at DEMO in 2008, Speaklike specializes in highly automated, rapid turnaround professional translation.
Speaklike, based in New York, was founded with the idea of professionally translating real-time communication, initially IM conversations. While the market for IM translation was initially limited, the company solved an important problem by creating a platform that is optimized for fast, high quality translation. This turned out to valuable in several important markets.
Their platform provides end-to-end automation of the translation process, from the initial translation request (delivered via a web API). Meanwhile, the translator work model emphasizes continuous work flow, contextual quality guidance, and improved productivity, combined with real-time interaction from end to end. Rather than work on large documents or projects, translators simply log in and see many smaller jobs, and can pick them from the queue as they work.
SpeakLike focuses on three major types of projects or media: CRM or customer support, web/news content and social media.
With CRM and customer support clients, they process small jobs 100-200 words in length, offer 2 to 4 hour turnaround (for email and support tickets), and for real-time communication, very small jobs (2-50 words) with a 15-30 second turnaround time. They also have to support many different languages, in both directions, and have integrated their service into popular CRM platforms, which eliminates project and translator management cost and delays. Their API based platform made system integration straightforward, and eliminates project management and manual work. This led to reduced handling, process flexibility, and improved quality, combined with 50% cost reduction, with average 2 hour response.
For news content, their challenge is to deal with large volumes of content, deliver consistent quality and style, with handle SLAs as short as 4 hours. In one project, they were tasks to translate 40,000 words in 24 hours across 9 languages, and were able to meet that deadline thanks to the automation of their platform. Most importantly, document preparation and project management automation reduced overall time required by 80-90%, and reduced costs by 50%.
An example of social media translation projects they are working on is Eurodis, a community for rare diseases in Europe. Users can request translations for discussions when needed. This is an interesting example of crowd-funded translation, where interested users pay to have posts translated, and can be applied to many domains such as online store listings, travel sites, and more.
Sandy predicted that over the next five years, we will see a merging of process automation, translation memory, and machine translation, a hybrid model to combine human and machine translation, and leverage human translation to continually improve machine translation quality.