Anita Joseph caught up with Samer Mohamad, Yasmina Regional Director for MENA at Yango, the global technology company that recently introduced Yasmina, the human-like AI assistant.
What inspired the development of Yasmina? Can you share the key features that set Yasmina apart in terms of cultural sensitivity and assistance with home, office, and day-to-day tasks?
At Yango, we believe in transforming technologies into everyday services for local enrichment. We created a smart assistant that is human-like, culturally intelligent and built specifically for the people of the Middle East. Yasmina understands major Arabic dialects and speaks GCC Arabic and English fluently.
The assistant also understands who is interacting with it, whether it is a child, a man, or a woman, and adjusts its answers accordingly. On a more practical level, Yasmina acts as a smart home center: it can adjust the room temperature, turn on the vacuum, or find a new recipe for dinner.
More than just an app on your phone, Yasmina is a human-like AI assistant that can be built into other apps and smart home devices, including smart speakers and home appliances.
How do you see Yasmina contributing to the broader landscape of smart assistants in the Middle East? What challenges and opportunities do you envision for the adoption of this type of technology in the region?
There are multiple ways smart assistants can continue to develop in the Middle East. First of all, I think we will be seeing more localisation from global companies. They will be making their products hyperlocal to meet the needs of specific regions and populations. Yasmina is already part of that trend. Being proficient in major Arabic dialects and having an intuitive understanding of Khaleeji culture, Yasmina represents a significant advancement in smart assistant technology.
Another big development we will be seeing more of in the future is the further integration of emotionally aware AI into our lives. At the moment, despite AI’s ability to swiftly deliver information, it often falls short when it comes to grasping human emotions, with UCL research indicating a 48% to 62% accuracy range.
We put a lot of effort into making Yasmina human-like and capable of maintaining conversations that feel like you are chatting with a real human. We believe this personalised experience will be appreciated by the community.
Speaking of personalisation, can you shed some more light on how Yasmina enhances and personalises user experiences?
As we set out to develop a smart assistant for the ME region, our mission was to build an assistant that would truly serve the sensibilities, preferences, and needs of the region.
As I have already mentioned, Yasmina distinguishes between children and adults, men, and women. Yasmina provides age-appropriate content when addressed by a child and uses correct grammar forms in Arabic depending on the gender of the person interacting with it. Yasmina will also give you culturally appropriate recipes. For example, it will not suggest you cook pork for dinner or serve wine to your guests. The assistant can also inform users of prayer times or play Adhan.
What challenges did you face when developing and introducing Yasmina? How do you think this milestone will influence smart assistant adoption in the region?
I think one of the biggest challenges was to make sure that Yasmina understands major Arabic dialects and responds seamlessly in Khaleeji Arabic. The Arabic language is extremely rich in both vocabulary and dialects, which vary from one region to another.
Yasmina’s ability to maintain an engaging conversation is based on a proprietary large language model. Dozens of copywriters and hundreds of individual content contributors edited and perfected dialogs to fine-tune the model, allowing it to grow into Yasmina’s unique Khaleeji personality. For Yasmina’s speech recognition, we worked with hundreds of residents from across the Kingdom of Saudi Arabia, including places like Riyadh, Jeddah, and Dammam. Their contribution helped us create an assistant that understands a wide variety of speech patterns, intonations, and dialects.
The smart assistant market is evolving. Looking ahead, AI and smart assistants will likely be even more attuned to various cultural and linguistic nuances. They will know multiple regional languages and dialects and understand local customs and social norms. For example, Arabic-speaking assistants will likely be able to consistently respond in the same dialect as the speaker, ensuring an even more customised experience.
Yasmina is designed to help users manage their homes, offices, and lives. Can you provide some examples of how Yasmina’s unique capabilities have made a tangible impact on users’ daily routines or productivity?
Yasmina is an enabler of a connected and more convenient life for the people in the Middle East. Yasmina is capable of transforming homes into smart homes, remembering homeowners’ preferences, and helping them with day-to-day chores. Yasmina can set reminders and alarms or adjust the room temperature according to your preference. The smart assistant speaks GCC Arabic and English and can easily switch between the two languages in mid conversation if that is what the user wants.
Yasmina signifies the new mainstream, which is all about making assistants more human-like through the advancement of language models and the careful insertion of the technology into the cultural context.