Despite Arabic being one of the most spoken languages of the world, and one of the six official languages of the UN, language technology for Arabic is many times way behind other European languages. Arabic language technology has not been able to profit from the resources available for many much smaller European languages, and there are very few tools available even for more fundamental NLP tasks such as segmentation. Arabic is a challenging language to tackle with its beautiful and peculiar scripting language, being RTL, the phenomenon of transliteration, its exceptional morphological richness, and last but not least all the spoken dialects. Considering the market of Arabic language users, of Modern Standard Arabic (MSA) or any of its dialects, there is a huge potential to break language barriers in the Arabic-speaking world and bring it closer to the rest of the world. This is what this presentation is going to be about. How a small LSP in the MENA region, Tarjama, boldly decided to tackle the Arabic language by building its proprietary language technology, including an NMT engine and a spell&grammar checker integrated with its TMS&CAT system. We will describe this exciting transformation journey of the business with its challenges and learnings and the impact it has had on the business becoming a pioneer in the region.
