Your accent can have a major impact on your job in customer service. There shouldn't be a worse or better accent. However, in today's global economy (though who knows what tomorrow might bring), it is important to sound American or British. Many people undergo accent neutralization training. Sanas, however, is a startup that has a $5.5M seed-round and uses speech recognition and synthesis in order to change accents of speakers in near real-time.
A machine learning algorithm was developed by the company to recognize speech and produce the same words locally, that is, without the use of the cloud.
It fits right into the OS' sound stack, so it can be used with almost any audio or video calling software. The company currently has a pilot program that includes thousands of participants from all over the world, including the USA, UK, India, Latin America, the Philippines, India, and other countries. By the end of this year, accents will be supported in American, British, Indian and Filipino languages.
Truth be told, I was initially bothered by the idea of Sanas. It was a concession to those who think their accent is superior and others are below them. Tech can accommodate the bigots. That's great!
While I still feel a bit of this feeling, I see that there is more to it. It is simpler to understand someone if they speak with an accent similar to yours. However, customer service and tech support are a large industry that is mostly performed by people from other countries. You can either place the responsibility of this basic disconnect on the entry-level worker or on technology. The difficulty in understanding oneself remains. An automated system allows it to be done easier and allows more people do their jobs.
This is not magic, as you can see in the clip. The character and cadence the person's voice are only partially retained and the end result is much more artificial sounding.
The technology is constantly improving, and, like any speech engine that's been used a lot, it only gets better. The American-accented version might be easier to understand for those who are not familiar with the accent of the speakers. This will likely result in better outcomes for the callers, including the person who is supporting them. Sanas explained that pilots are still in training so no numbers are yet available. However, testing has shown a significant reduction in error rates and an increase call efficiency.
It is good enough to attract a $5.5M seedround, which includes participation from Human Capital and General Catalyst.
Maxim Serebryakov, CEO of Sanas, stated that Sanas aims to facilitate communication and remove friction so people can communicate confidently with each other no matter where they are or who they are communicating with. This mission is hard to argue with.
Although the ethical and cultural issues of accents and power differentials will never go away, Sanas is now trying something new that could be a powerful tool to communicate with professionals whose speech patterns can hinder their ability to do so. This is a worthwhile approach to explore and discuss, even if we could all just understand each other better.
