Our Digital & Social Media Coordinator Lucy Woods speaks with Neil MacDonald who recently joined Speechmatics as Chief Revenue Officer. Having recently announced Speechmatics to feature as a finalist in the Fast and Furious Pitch Off (press release here) at Venturefest East, we wanted to find out more about this rising start-up.
What is Speechmatics and how was the idea borne?
Speechmatics is the culmination of 3 decades of advanced Machine Learning and Artificial Intelligence research conducted by our founder Dr Tony Robinson. We apply the latest research in Neural Networks to the task of understanding speech. We provide an automatic speech recognition technology that unlocks the data contained within what we say to each other verbally.
How has the idea been funded?
We have been self-funded thus far, and are already cash-flow positive. But we are now looking to accelerate our go-to- market strategy by seeking investment, enabling us to accelerate development of new languages, for which we built a ground-breaking new approach. Also, we will be continuing to push the boundaries in speed and accuracy, always staying ahead of our competition.
What’s your association with Cantab?
Many of our team studied at Cambridge University and as a company we keep a very close relationship with the University.
What stage is Speechmatics at right now?
We are in “growth and acceleration” mode. We have grown the company organically thus far, without any fundraising. However, given the huge market opportunity we are now considering raising growth capital
Tell us a bit about the team behind Speechmatics?
(How big is the core team? We see you’re on a recruitment drive at the moment – is hiring a challenge?)
We are now 18 full-time employees with 6 PhD’s and are growing at a very rapid rate. We find that the quality and pedigree of our team acts as an artificial gravity drawing on the brightest new talent to the industry. And at the same time, we are adding seniority to the team to ensure the best efforts of our R&D are maximised to their fullest potential.
Finding candidates with the right skills is of course challenging, given the draw to the large US mega-tech companies, but we believe that a young Cambridge start-up such as Speechmatics offers unique and enjoyable challenges that cannot be found with these largescale American bodyshops. Culturally, we provide an environment that ensures our colleagues work on tasks that interest them personally as well as professionally, and we see ourselves really as the archetypal garage-based startup…I don’t think anyone actually owns a suit and tie here.
Tell us more about the technology being used
Our technology has been developed over many years of ground-breaking leading-edge research, led by Dr Tony Robinson. Based in the disciplines of Artificial Intelligence and Machine Learning, our solution to the difficult problem of understanding speech is to model the methods used in our own human brains; similar to a child growing-up, we train our system on billions of words and audio and over time it improves.
You can test the software for free at www.speechmatics.com
In terms of business development & getting the Speechmatics name out there, what are you doing to achieve this?
Speechmatics are firm believers in partnership and cooperation with complementary technologies; we do not exist in a vacuum. Networking events allow us to expand our sphere of influence and provide us the means to be recognised thought-leaders in our discipline; to be the name on the tip of the tongue when people speak about speech recognition and artificial intelligence.
Let’s talk competition. What other speech recognition software is available?
There are a few companies in our space, but even fewer use deep learning in their product. Also, we have a very strong research ethos and capability to deploy the very latest thinking in how to solve the speech recognition problem. The industry is dominated by a single player, Nuance, whose cadence for innovation cannot match ours.
No where on the website is cost mentioned (not that I can see?) what are the options available? i.e. is it licensed or one of payment?
We do not dictate how our customers do business with us, recognising that every customer and industry is different. We will adapt our business model to dove-tail with the model of customers, creating a win/win situation.
Whats the plan for the future?
Continue to innovate and move toward a point where computers can understand and interpret every word we say. We are also developing something, which we hope, will completely change the way people interact with machines.