The global banking crash of 2008 highlighted how easy it was for markets to be manipulated by traders using tools provided by their employers for communications, such as email, chat and telephone, and how little effort was being put into monitoring those channels. I had already been working in the field of Natural Language Processing for text and it seemed logical to extend that work out to voice as well.
Tell us about the business – what it is, what it aims to achieve, who you work with, how you reach customers and so on?
Our core focus is on understanding human interactions, particularly in the area of speech. Our core focus is on areas that other companies shy away from, which is highly secure communications, such as banking, where it is important that none of the data is ever sent to the cloud, but is under the complete control of the bank itself. We pioneered the use of GPU cards for speech recognition back in 2014, and since then have expanded our portfolio of speech tools to embrace areas such as identity management, summarisation, call wrap up notes, and credibility analysis for the insurance fraud market.
We work with many of the world’s largest banks, mainly via a network of partners spread across the globe. Like all companies, we started by selling direct, because if you can’t sell your product, you can’t expect other people to, but our success meant that others could see how our software embedded in their proposition could be a win for both parties. We still do a lot of direct marketing, mainly via social, but it is for brand awareness, rather than attempting to sell our software directly.
How has the business evolved since its launch? When was this?
The business first launched in 2009. It took a while for the banking world to be ready for speech processing technology on the trade floor, so we have been through a number of iterations. We established ourselves as the leading provider of speech solutions to the legal eDiscovery industry, having cut our teeth on a lot of LIBOR-related material in the early days, and worked in other fields such as mobile phone and meeting room recording as well as an augmented reality meeting room product which was several years ahead of its time. In recent years, we have put a lot of effort into helping reduce fraud on telephone claims in the insurance market, an issue which climbs rapidly in times of economic distress and uncertainty.
Tell us about the working culture at Intelligent Voice
We are now fully remote, so that means we have complete trust in our team to get on with their jobs without interference. I love the passion in the team for what we do. The company is strongly R&D-based, so there is always a lot of cutting-edge development to be done, and that gives a great deal of job satisfaction, especially when that work ends up being used by some of the world’s largest banks. We recently had to empty our offices and move our entire data room in very short order when our landlord got into financial difficulties, and the entire team swung into action with no complaints, and did the move in under 3 weeks from a standing start. And all of this while we were still delivering support and software to customers. We do try to get the team together for social events, and make sure we have an annual kickoff to set the scene for the year. This time round we handed out a couple of “Unsung Heroes” awards, as there are always some people who don’t get the recognition they deserve, and I wanted to show them that we did appreciate what they did quietly in the background.
How are you funded?
We are almost completely self-funded. We did take in some investment a couple of years ago as part of a larger European cloud play that was going to take on some of the large US AI giants with a focus on sustainability. It was a great idea, but couldn’t attract the necessary investment for the wider concept.
What has been your biggest challenge so far and how have you overcome this?
That depends on which day you ask the question! If I look back, it was probably getting the first couple of customers. I had thought it would just be getting the first customer, but in actuality, we needed to get the roll going, to demonstrate that it was more than a fluke that someone wanted to buy the solution. Selling without references is very hard, because no-one wants to be first (or second), so we had to be creative with pricing, as well as going through endless trials and PoCs to prove out that the solution would so what we asked of it.
How does Intelligent Voice answer an unmet need?
Not all businesses want their data in someone else’s cloud. Yes, sending data through an API is convenient and easy, but you often can’t guarantee where it will end up and what people will do with it. So many companies have quietly changed their privacy policies to allow them to use customer data to train their own AI models. We have always has a security-first approach, which means allowing customers to have cloud power and API convenience, but running in their own cloud or data centre. As we expand the portfolio, this means that our customers can not only run speech processing securely, but now LLMs, as we have introduced them into our pipeline.
What’s in store for the future?
“AI” is the buzz word of the moment, but it is going to transform a lot of industries, and we have been at the forefront of work in this area for many years: We were building Explainable AI systems using Transformer models back in 2019. The focus will be on delivering this technology in a secure and cost effective way to our customers with use cases that make sense to them. We've also been working on some more advanced features, like self-learning speech recognition and encrypted inferencing, and these will see the light of day early next year.
What one piece of advice would you give other founders or future founders?
Make sure you have enough cash, for you and your business, for much longer than you thought you would need it. Cash will always be King in running a business.
And finally, a more personal question! What’s your daily routine and the rules you’re living by at the moment?
I tend to start early, around 7.30 or 8am, depending on whether it is a cycling day or not. I stop at 12 on the dot, have lunch and go for a walk to clear my head, and then back to work. If I can, I try to get into London one day a week: We recently lost our office when our landlord imploded, and we still haven’t decided whether to stay fully virtual (we were pretty much there before). I like to see people, so go out of my way to try to organise things face-to-face.
As soon as an email comes in, I am tempted to answer it, but I am trying to flag things to deal with later, sometimes because I just don’t have the time, but also because I find I need to reflect a bit more. Too many people see an email (or chat message), and type the first thing that comes into their head: And that can be more emotional and less well thought out than it should be.
I really believe in treating people the way I would want to be treated, having had some shocking bosses in my pre-entrepreneur days. And that works out for the most part. Some people over the years have frustrated the hell out of me, but that is usually because they are not doing their jobs when everyone around them is, and I feel bad for the others who are. It’s even worse when they just can’t see it, and the effect they have on the rest of the team. We have built up a great bunch of people at IV, and that is worth fighting for
Nigel Cannings is the cofounder and CEO of Intelligent Voice.
