Understanding Every Voice: The Aim of Speechmatics

At Speechmatics, our aim is to make voice technology inclusive and accessible to everyone in the world, regardless of demographic, age, gender, accent, dialect, or location. In its simplest terms, we want to understand every voice. While we have a focus on how our technology gets us to this goal, something we must always consider is the human angle.

Our industry-leading Autonomous Speech Recognition (ASR) is pioneering for several reasons, not least because of its ability to understand the widest range of dialects. We’ve developed a Global speech-to-text language pack for English and Spanish, which means the technology can understand all accents and dialects within the language, negating the need for more complicated and less effective multi-layered systems.

It’s this people-first thinking that puts us ahead of our competitors. Understanding every voice must be led by the people. A core value that must be embodied to be believed. At Speechmatics, we’re continually looking at how we can implement understanding every voice into every aspect of what we do.

Our First Steps

As you’d expect, understanding every voice takes time. The tech is still evolving, and the company still growing. Naturally, it wasn’t always like this. The start of Speechmatics’ growth journey included – ‘Live for the Wow’, ‘Build Authentic Relationships’, and ‘Be the Adventure’.

While a good start, these values don’t reflect our business as it is today. That’s why we have new ones:

• People First: hiring the best candidate for the role, not necessarily the most qualified.

• Move Fast: Empowering employees to “fail fast and move on.” Creating an environment where we can “debate freely, make timely decisions, and commit to outcomes.”

• Care Deeply: This is where the ‘understanding every voice’ aim originates from. We want to ensure we put people first, especially when it comes to the impact our actions have on the world.

• Be Ambitious: Attempting to make change within the industry through big goals and breakthroughs.

Understanding every voice is the glue that holds our present values together. Plus, they’re much more relatable for new recruits giving us a better chance of hiring the best person for a role – something essential as we grow.

Shaping the Company

Speechmatics’ growth, while rapid, mirrors human life. Starting out, we had one aim: to grow. But soon we’re hitting the business equivalent of puberty, which means it’s time we take a thorough look at our path ahead. Much like real life, our aims, and values ebbed and flowed over our development.

The company was very much tech and machine learning-focused while still having that ‘people first’ culture. Over five years later, we are getting close to 200 employees. The more Speechlings there are, the more aspects of human life we need to consider. Nowadays, there is much more of an onus on everyone at the company to take our global approach to ASR and embed it in the hiring process.

If it’s people’s needs leading our direction, we must understand them.

Our Plans to Truly Understand Every Voice

At this moment in time, our ASR can transcribe 31 languages. While there is still some way to go, we are constantly looking to expand. Language, after all, is universal. One of the best things about this is we do not need a retrospective, hindsight-laden analysis to fully comprehend the power of understanding every voice.

We see it every day in the work we do, in the data we collect, and in the products we release. It’s this inclusivity we are looking to push even further.

After all, change happens one word at a time.