Charles Corfield is president and CEO of nVoq of Boulder, CO.
Tell me about yourself and the company.
I started off life as a mathematician. The company that we’re talking about today is nVoq, which is based in Boulder, Colorado. It is a tech company. It does voice-assisted workflow, or voice-automated workflows.
Did you find that you had an aptitude for entrepreneurship that you didn’t expect, or is it truly related to your mathematical training?
If I may quote someone, when I was a teenager, I had the good fortune to observe a couple of entrepreneurs in action in the UK. John McNulty was his name, one of those entrepreneurs, who said, “The reason most people become entrepreneurs is because they’re fundamentally unemployable, so they have no choice.” [laughs]
If you’re one of these people who is a constructive troublemaker, that you’re always poking at things and asking questions, then entrepreneurship is a fairly natural thing to do, even if by personality traits you may not be the most obvious candidate for it.
For example, mathematicians are notorious for being somewhat shy, retiring, and socially awkward. I certainly plead guilty to being something of a social retard myself. The joke used to be that you can tell an extrovert mathematician because he or she stares at your shoelaces instead of his or her own. [laughs]
What you figure out as a mathematician is that there’s a pattern to all these things. Go learn the pattern and go figure it out. It’s really actually not that strange for a mathematician to become an entrepreneur.
What characteristics about yourself, other than being rebellious, would you say have been important in your career as an innovator, investor and now running nVoq?
At least willing to ask awkward questions, if you can characterize rebelliousness somewhat more charitably.
Two attributes which I think are key in this are, one, the willingness to sink your teeth into something and just stick at it. One of the things that I have observed over the years as I applied my trade in technology is that many people have folded their hands far too early. They’ve just sort of given up. Somehow they didn’t in the end have the courage of their convictions.
That brings me to the second point, which is, mathematicians can often have insights into the way things work and see things which are not always easy for other people to see. If you have the good fortune to have the right insights, then that’s probably more important than having a big VC backing you. In other words, good insights can make up for a shortage of dollars.
You were an investor in BeVocal that was sold to Nuance a few years back for a pretty good chunk of change. That became Siri, right?
That I could not comment on. [laughs]
I wondered if I’d get a “yes” out of you on that.
I think I shall refer you to Nuance to comment on matters of Siri or otherwise. [laughs]
How did you get involved with speech recognition?
The story behind it was that I met up with several would-be entrepreneurs in Silicon Valley who wanted to do something with speech recognition. They were not ready for prime time as far the VC community there was concerned.
However, I liked what I saw, and so I worked with them in formulating the business. I invested in at as well, as did eventually a number of VCs once they got to a stage where they were a candidate for taking funding from the VC.
It was an interesting model, because before we had the term "cloud," they were actually doing a cloud-based IVR. This was also one of the not very common times when you could do a gain share model and control enough of the levers to make it work.
In that environment, it was well known in the industry what percentage of your incoming phone calls to customer care you could automate, or not as the case may be. If you couldn’t automate it, it had to go to an agent and that’s really what drove your expense. The approach at BeVocal was that we would use a judicious amount of speech recognition to increase the — as some people call it — call deflection, meaning deflection away from an agent, or call automation or containment within the IVR.
The deal we would make with the customer is that for every percentage point we can increase that automation, you pay us X cents per minute. That turned into a very good business model. The reason that type of model is not very common is because often technology companies can’t control enough of the levers to influence the outcome in their favor. Gain share models are often very good for the client and lousy for the technology company.
Nuance is probably the name people people think of most often when they hear the term speech recognition. How are NVoq’s offerings different and how do you compete against Nuance?
We take a different approach. As you said, Nuance is the brand name or the 800-pound gorilla that is known in healthcare. Their primary offerings are back-end transcription as they have absorbed transcription companies and put that on to their back-end speech recognition. Then the front-end product, hich is more widely known, the medical version of Dragon. That is a desktop product. It’s what is called a fat client. All the functionality has to be installed either on the enterprise server or the user’s desktop.
Our observation is that by taking a different approach, which is to supply functionality out in the cloud, we are able to meet the needs of people who are more cost conscious and need a very simple and portable access to speech recognition. By simple, meaning it’s very easy for them to learn what they need to learn. By portable, it respects the fact that they are working in multiple locations. They’re going from offices to clinics to hospitals and so on and they really need just one account that can follow them around.
The cloud, as long as they have Internet connectivity, allows them to hook up to their account wherever they are. Then from a user experience point of view, what we have focused on is to make that process upon boarding the user — that is, training them up from ground zero — very simple for the user. The process of supporting that user in their daily use is to make that very simple as well.
Let me give you a for instance. Because the functionality is in the cloud, we or the reseller can see exactly what the user is doing during the early days and can make judicious interventions to true things up for that user: introduce vocabulary items or tweak the system in a way that meets the user’s actual usage. What is nice for the user is that the system seems to be proactively addressing their needs without them having to pick up a phone and ask for help.
This brings us to, I think, one of the big opportunities of using speech recognition in the healthcare space, which is to get a higher adoption rate. Nuance has in effect set the standard, so you will see roughly 50 percent of people who have started on Dragon end up abandoning it. Not because Dragon is a bad product. Dragon is a perfectly good speech recognition product. The issue is that when they need support, it’s not convenient to get it.
We make a very strong push in that direction of delivering good customer service and timely customer service that makes the difference for these users. Because to be blunt, they’re all far too busy to pick up manuals on speech recognition or wade their way through indexes trying to figure out, what did I get wrong? Why isn’t this working for me? Far better that before they even realize they’re having issues, someone can intervene behind the scenes and make the system do what it needs to do.
How do you see the market for voice-operated commands in healthcare or the use of speech recognition by non-physician clinicians for something other than dictation?
If you consider the numbers, there are 800,000 physicians, plus or minus, in America. But the total number of people working provider side in healthcare is closer to 16 million. There is clearly a large, unserved market or potential market of people who need something which can speak to their needs, speak to their workflows, if you will. It’s simple. It’s affordable. It can automate their rote tasks.
Providing a solution for these people is something we are very interested in and are already doing. We look at it as being ultimately that we should see millions of people who are working on the provider side who are able to benefit from driving the EHR or whatever application they’re using for scheduling or some other type of documentation where they can use voice where appropriate.
I don’t mean to ask too many Nuance questions, but companies that have been successful in anything vaguely related to speech recognition usually end up being bought by Nuance. Is that a concern of clients or an interest that you have?
Well, the future’s always very hard to predict, isn’t it? So I shall defer on that one. We’ll stay focused on providing a very attractive user experience and also financial experience for the users. Where that takes us in the future, who’s to know? [laughs] We’re not courting Nuance, nor are they courting us.
Talking about those potential non-physician users, how do they find you or how do you make your presence known in ways for something the average hospital hasn’t thought of?
There’s nothing like word of mouth that you make something easy for someone who had no idea it’s possible. The fact is that Nuance has invested heavily in creating awareness of speech recognition. So people have thought about potential applications, but they may not be able to implement those applications using what’s available from Nuance.
As much as anything, that’s just a fact of life. It’s very hard for one company to cover all possible eventualities. We focus on the ones which are probably not in their sweet spots. But we are in a sense down market from where they are pushing with natural language recognition, the coding engines and what have you. We are much more focused on bread and butter and workflow, and in a sense, a more mass market offering.
I don’t know how you distribute your product or who your customers are, but who’s doing something really interesting with it that would be a notable name?
First of all, how people are getting their hands on the product. The approach we take is it’s channel based. We will work people in the reseller community who, over the years, they know a lot about end users in their neck of the woods. They know where to go hunt, so to speak.
I think in respect to people whether or not they want their names used, we do have end users who are some well-known names and who certainly appreciate the fact that there is an vendor out there who is taking an attractive approach both for support and also financially. Budgets are under pressure and it’s a very low-risk way for them to use speech in their applications, because for example, we are a subscription base, which means the financial risk is fairly low. If you really don’t like the product or it doesn’t work for you, well, stop paying. [laughs] It’s a monthly subscription, as simple as that. On the other hand, if it works for you, the fact that it’s now a monthly expense rather than a large capital outlay is for a number of users a very attractive proposition.
Other than BeVocal, one of the other big successes you had business-wise was Frame Technology. You sold that to Adobe for $500 million a while back. You’ve had a lot of success in creating and selling these companies. What kinds of investments would you be looking for today in healthcare?
Everything around workflow. There’s opportunity here to look at a script we have seen before, which is with the ERP software or database software that took place in the enterprise world. You had companies like Oracle and SAP and Powersoft and others rising out of that technology wave, if you will.
The big databases are in a sense the equivalent of the big EHR systems going in. Now that we are probably most of the way through adoption of EHR, that big data repository is now in place into the hospitals or clinics. The opportunity is now for a second generation of applications to come along which can ride on top of the big iron EMR and they can then address particular types of workflow.
I think we will see a wave of companies emerging in the next five years who build on top of the EHR and go and address some of these point workflows that are hard for the big manufacturers to address because they already have their hands full with Meaningful Use and a list a mile long from their clients about the other things they need.
What are you priorities or strategies for the company for the next few years?
It’s really all about customer service. We are in the business of productivity, taking cycles out of people’s workflow. Anywhere where we see inefficiencies that we can address, we go after that.
The thesis in high tech is that it’s really an arbitrage game if you will, because you’re always taking an existing process and re-implementing it, leveraging technology to lower the cost point of that process. The difference you’ve opened up between what it costs today versus what it will cost once you put in the technology – that’s the arbitrage that you can then take your cut of and run a business on. So for us, it’s all about productivity.
Do you have any final thoughts?
For anyone reading this interview, if you would like a very friendly and approachable and high-impact customer service approach to using voice recognition in a workflow, come give us a call. I’m sure we can make you happy.