California-Based Kanari AI's Arabic Speech Technology Solution Is Looking To Tackle The Communication Challenges Posed By Different Dialects Kanari AI's speech technology can be adopted across a variety of sectors, ranging from media and education, to legal and government, and it also provides valuable assistance to firms deploying security and voice assistant services.

There are over 20 Arab countries in the MENA region, and with each country having its own set of distinct dialects, it poses a communication challenge that often goes under the radar in many industries.

To cater to this specific issue, California-based startup Kanari AI launched its dialectal Arabic speech technology solutions, with specific features that focus on automatic speech recognition as well as text-to-speech conversion, in 2020. "There is a lack of speech technology adoption in the MENA region, and our technology addresses the four main pain points of speech technologies in the Arab world: poor accuracy, deployment concerns, lack of customization, and complicated integration," explains Kanari AI founder and CEO Ryan Carmichael. "We detect 19 Arabic dialects in one global speech model. We have also licensed our technology from Qatar Computing Research Institute, and have since improved accuracy and focused on commercialization efforts."

According to Carmichael, Kanari AI's speech technology can be adopted across a variety of sectors, ranging from media and education, to legal and government, and it also provides valuable assistance to firms deploying security and voice assistant services. But detecting a variety of Arabic dialects isn't all that there is to the startup either.

Within its automatic speech recognition offerings, it can also use pre-existing datasets, that Kanari AI's clients have on their own systems, in order to increase the accuracy of its results. It also enables on-premise, cloud, and hybrid cloud solutions that make its speech-to-text features flexible and quick. The team behind creating this technology is spread across the globe, which makes the startup what Carmichael calls a "global remote company.

The Kanari AI team currently comprises of 10 members, with the founder himself based in California, while the engineering team is currently located in the Indian city of Bangalore. "Our Chief Scientist is based in Doha, and we have artificial intelligence (AI) researchers in Lebanon, UAE, and Egypt," adds Carmichael. But while having such a diverse, remotely-based team poses a challenge in itself, Carmichael also had to deal with the problems associated with launching a startup in the middle of a global pandemic. "Launching a global startup during COVID-19 has been challenging, especially with in-person relationship building being important for enterprise solutions," he says. "But having a local presence in Dubai has allowed us to not only build our local brand, but has also provided more access to top AI and engineering talent."

Joining the Mohammed Bin Rashid Innovation Fund (MBRIF) accelerator program also helped the company in this regard. "MBRIF was able to provide much needed regional support for our company," adds Carmichael. "The MBRIF team has been very responsive by helping us in setting up our local presence, and they also introduced us to enterprise partners and local investors." Given the increasing number of industries in the region that are finding themselves in need of Arabic speech technology solutions, Kanari AI looks to be all set to capitalize on the opportunities ahead- it's only a matter of time.

Related: Abu Dhabi-Based Giisty Is On A Mission To Restructure How Organizational Communication Works