This site uses cookies to provide you with a more responsive and personalised service. By using this site you agree to our use of cookies. Please read our privacy statement for more information on the cookies we use and how to delete or block them.

Talking tech: Interview with James Duez, CEO and co-founder of Rainbird


June 2020
Read time: 9 minutes

James Duez, technologist, entrepreneur and investor, is CEO and co founder of Rainbird, an intelligent decision engine. He has over 25 years’ experience building companies, with a focus on tech and more traditional companies looking to transform their growth through exponential technology adoption.

 

He has worked with 250 Global organisations, governments and investors, and is a member of Abundance 360, a 25-year program led by Peter Diamandis to leverage exponential technology.

We caught up James to discuss knowledge mapping, Bagpuss, an essential NHS project, and why start-ups shouldn’t see themselves as businesses…


So, James, tell us something about your own journey. What was the path that led you to Rainbird?

I was originally a software engineer back in the late 1980s and early 1990s. I came to the conclusion that corporate life was not for me in about 1993, and decided to set up my first business, a software company. In the process, I gradually transitioned from the technology side to the business side.

I ended up co-founding a succession of businesses in a whole range of areas but largely focused for 10-15 years on compliance, risk, information security, data security. I did projects with all sorts of organisations from the Pentagon to most of the world’s banks – trying to help them keep the right side of regulators.

I sold that business in 2005 to a public company and then ended up doing M&A for a few years, which taught me a little bit about acquiring other businesses. In 2007, I became an angel investor, which was really my way of finding the next adventure. After investing in a number of businesses, I joined one in the insurance space called Validus IVC, which was a good link into Rainbird because Validus was a business that had identified a huge amount of fraud and inflation in motor insurance claims.

There was an idea to take a couple of experts that knew how insurers were being manipulated, and build an insurance tech business that defended this on behalf of insurers. I thought this was a great investment but I didn’t understand why the plan was to start a consultancy business. I said, ‘You don't want to start a consultancy business. You want to start a software business,’ and my co-investors said: ‘Well, we don't know how to do that. Would you like to come and help us?’

And that business actually ended up revealing that 85% of all the third-party motor claims in the country were being manipulated systematically. We ended up with our case on Panorama and giving evidence to Jack Straw, who went on to change legislation around some of the things that we found. So that was a really interesting journey, but perhaps the most fascinating thing about it is was that we had a software system which at its peak saw something like 25-30% of all UK claims and worked for 26 out of 30 insurers… but was all actually just based on what a couple of smart people knew. We took that knowledge and embedded it in software so that it could be wielded at scale. The business was very successful and we sold it to private equity in 2013. I went back into investing and eventually ended up co-founding Rainbird.

And there was a similar idea behind Rainbird?

Yes. That earlier business had required 25 or 30 engineers to work for nearly two years to build a system. So we thought: Wouldn’t it be great if the world had a platform that would enable you to do that in a month or two? And that’s where Rainbird came from – it’s about the shared experiences of trying to scale human knowledge in software.

It’s really interesting that you have both the technology and the business skillsets in your makeup. Does that make you quite unusual?

I’m basically a software engineer turned businessman turned investor and repeat founder. I love start-ups. I love getting in early on something and trying to overcome that inertia, build something that the world doesn't have.

It’s true that the two sides don’t always go hand in hand, but I've always been fascinated by the challenge of taking technology that is actually understood but not necessarily fully commercialised. My career has been standing at the path between businesspeople or investors, and technology people. I’ve somehow managed to develop a set of skills that are not left-sided or right-sided. I’m not auditory. I’m not visual. I'm a mixture of all of those, and I think entrepreneurs have to be a Jack-of-all-trades. We’re not masters of anything. We need to be able to do 80% of what you need to know about marketing; about people; about tech; about sales – and then blend that together in an organisation to find value, and then pass that on to a specialist that knows more than you do, and then eventually go on and do something else.

That’s what I've always got a thrill out of angel investing: for me, it’s about trying to find great teams with great tech that I think have the potential to be globally significant but maybe haven't realised that yet.


About Rainbird Technologies

Industry; Computer Software
HQ: Norwich, Norfolk
Ownership: Privately Held
Size: 11-50 employees
Visit website


So where does the name Rainbird come from?

It goes back to an old kids’ TV programme called Bagpuss. There was a character called Professor Yaffle, who was this wooden woodpecker. Professor Yaffle was based on the philosopher and mathematician Bertrand Russell, who is now seen as one of the forefathers of AI. He and Alfred Whitehead wrote a book called Principia Mathematica, which was the first time anybody had written down a formal method of mathematics. Without that, there would have been no Enigma machine, and AI wouldn’t exist as a field today. So the name is a skewed nod to Russell: Yaffle is an old English word for woodpecker – and so is Rainbird!

And how did Rainbird develop?

Ben Taylor, my Rainbird co-founder, was an ex-computer scientist at Adobe where he worked on the Acrobat and PDF solutions for them. He was my chief engineer in that insurance tech business, and when we sold the business he very soon left and approached me. He said: ‘I have this idea for a platform. I think we can build something the world doesn’t have. Would you like to take some of that money you’ve just made and invest in me?’ It took me all of 20 minutes to agree and we actually ended up founding the business together.

And what does it do?

Rainbird is that next generation of intelligent automation that can automate human decisions.

We would classify it as an intelligent automation technology that takes human knowledge and turns that into machine intelligence. People often think of AI as building models from data, but actually machine learning is not the only form of machine intelligence. We’ve been trying to build what we already know about the world into software for years using simple decision trees or business process frameworks.

Rainbird is a way that you can take all the nuggets of wisdom that an expert has in a particular domain and turn that into a knowledge map rather than a decision tree. This is a very holistic representation of knowledge that you can query, because it knows all the pieces of a particular subject. A typical use case would be in tax. If you want to evaluate an R&D tax claim or make an IR35 decision, you might rely on a human expert who knows all the bits and pieces of how individual decisions are made, and who you rely on to stitch that together and make judgements. Well, Rainbird can be used to replicate or synthesise that process at scale.

"You tell Rainbird all the bits of what you know. You ask it a question, and Rainbird can join all of that up. You can connect it to data or, if it doesn't have enough data, it can ask human questions at runtime when you query it."


So sometimes it presents a little bit like a chat bot. You ask Rainbird a question like ‘What's my IR35 status?’ or ‘Should we lend this person money?’ What Rainbird does with that is ask questions and at the end of that process it makes a judgement. Crucially, it can also explain how that judgement was reached, which is essential for decision-making in regulated industries. Rather than a probabilistic assessment, which you get from machine learning and is not always obviously interpretable, you get a definite answer, supported by an explicit chain of reasoning based on human inputs. People will trust a machine if the machine can explain itself in human terms, and that's the special sauce really.

Is it difficult to use?

You don’t need to be a software engineer to use Rainbird. You need to be reasonably technical but actually anybody can do this. We’ve had tax partners build models in Rainbird. It’s actually a very visual model, a bit like mind-mapping. You’re starting to put down the concepts and relationships that build a map of your world, the individual rules for how you make decisions. But you’re not chaining them together to make a decision tree – you're putting them on a map so that Rainbird can dynamically join them up to solve problems. A decision tree is always built to answer one question, but you can build a Rainbird model that can answer 200 different questions – including questions you didn’t even know you needed to ask at the time!

Rainbird is able to mash together the things we have in our head – all the rules we have accumulated through our formal education and our training, and the rules we work by. We ask questions. We gather data and we make judgements. That’s the process of inference. We do that something like 30,000 times a day. Rainbird can replicate that process with all of that uncertainty, and it all starts with a human initially building a model based on the area you want to focus on. So if you’re building a model in credit card fraud you might have a fraud expert build a model that contains every individual bit of logic that they've seen in their career around fraud. So then what you have is you have a machine that can do that at massive scale with millions of transactions an hour applying the very nuanced seasoned expertise of a very seasoned expert.

So it’s a way of unlocking the tacit knowledge that may be trapped in siloes across an organisation?

Every business that aspires to be digital now has to be digital as a matter of survival, and now people are more dispersed, there is no longer the ability to provide physical supervision. We need to use technology. Okay, it's great that we've got the VPNs and the laptops and Zoom and everything else, but we actually also need to be able to bring quality to the way that individual workers make decisions in their homes.

We need to bring efficiency in a world where there are huge pressures on costs, but in a way that’s human-centric. Our projects are humans being augmented by a technology that makes them typically a hundred times faster and the quality of their decisions, even at the top expert level, can be as much as 25% more accurate or better quality decisions, and two magnitudes faster.

Tell us about your NHS work

When it became clear that COVID-19 was going to change everything, we wrote an open letter inviting the world to use our technology for free where they felt it could benefit others in the face of COVID-19. And a number of organisations responded, including the NHS. 

There was some ambiguity around the self-isolation rules for staff – one hospital alone was having in excess of 500 calls from staff in a day not sure whether to come to work or not. It was paralysing the occupational health department. We took all of the government logic and the NHS logic – which, interestingly, didn’t quite align – and in eight days, we built a solution that was able to identify which members of staff or key workers should go to work or stay home.

The model has now evolved into a platform that is identifying who is a priority for testing. So if you're an NHS worker and you're perfectly okay but your child is sick with symptoms of COVID-19, the NHS will want to test your child because they may be on seven days’ isolation but you're on 14 days isolation and they need you back in hospital. You’ve got hundreds of people every day consulting with Rainbird about their symptoms, from which it can present to the hospital a prioritised list of who should be tested. And in a world where there have been limited supply of tests that’s really important. 

We provided this free to one NHS Trust, and more Trusts are taking it on. It’s now evolved into a project that's doing a much more comprehensive risk assessment of key workers. Rainbird can literally interview every member of staff, and tell them what zones they are safe to work in or whether in fact they should be working at home. It gives the hospital management the ability to know who can work where and when, and it can be updated as the guidance evolves too. 

"We've been talking about identifying PPE risks as well, and moving into mental health for frontline workers, ensuring they’re properly signposted to support services. And of course, a lot of this logic that we’re supporting the NHS with is applicable to other businesses that are also on the frontline."


For more on this, see our latest insight on How tech firms are doing their bit in the battle against COVID-19. It’s all about trying to make a difference at a time when the world needs to be using technology to solve some of these problems because of (a) the global scale and (b) the pace at which our knowledge around this disease is changing, which needs to be incorporated into decision making.

What’s the size and makeup of the business now? And what role has investment funding played in your development?

We’re eight years old. The business was originally accelerated by Techstars, so it had the benefit of going through one of the two biggest tech accelerator programmes, which is about trying to get two years’ growth into 13 weeks. We’re currently about 45 people. Ordinarily, our offices would be in Norwich and London, in Shoreditch, although obviously they are closed at the moment. We were always set up to work from home so for us there were no technical challenges for us doing that.

We have very supportive investors. I was actually the first investor in myself, and after that a lot of the investors were informal syndicates that have invested with me elsewhere. We’ve gone on to have some institutional investors too, and obviously that’s been key. We’ve raised about £10 million on our journey. We are revenue-generating; we’re very focused on growth and internationalising particularly, into North America. We’re actually very well-capitalised as well. We raised money just before COVID-19 and that’s important, because any tech company that is not revenue-generating or profitable and not well-capitalised is in peril.

We’re also partner obsessed. We’re not about going it alone. We work with five out of the top six global accounting firms and some of those are also partners because actually we believe that what the world needs now, more than ever, is organisations working together not competing. That’s always been in our DNA. For example, you take Rainbird’s technology, and you take what BDO knows about tax. Together, you have something very powerful. You have a two-plus-two-equals-more-than-four situation, and Rainbird as a technology can do that in any domain. It doesn’t care whether it’s healthcare or tax or fraud or credit or whatever.

How do you balance the need to keep innovating with the pressure to meet targets every day? 

Innovation is our lifeblood. We are very technically ahead – we’re in a position where we have a platform that is banking grade and very mature and stable: there's not been a blip in any of our live environments for over three years. We’ve also built another platform that’s close to coming to market, called Evelyn. Rainbird is about taking human knowledge and turning it into machine intelligence; Evelyn is about making machine learning explainable.

I think there are three things that are essential to being a start-up or a scale-up like Rainbird. Number one is you need to have scale-through partners. You need to demonstrate 100% year-on-year growth minimum. And you need a relentless investment in innovation in your technology and the patents that come with it. 

Internally, we have a team who are very focused on the existing platform and an R&D team who are developing new tech, and then we have a method that stitches those together. After that, we have a client success team which includes a big group of people who are experts at using our own technology – Rainbird ninjas that can help clients and partners build impactful things. We have Rainbird University, our training platform, and we run virtual courses so people anywhere in the world can have classroom training online and learn how to use Rainbird.

What would you say is the biggest challenge to scaling up a fast growth tech business?

I think the biggest challenge to scaling up in a business like ours is trying to help people understand that the problem you solve is even solvable. When you’re doing something that is a new technology in a new category, that’s very difficult. But ultimately, it's about making sure you find product-market fit and that you don’t try to scale until you have. And actually, for us the challenge is always ensuring that we adequately resource ourselves in marketing in order to have the reach and the voice we need to match the power of the technology.

Imagine a young technologist in a garage, dreaming of taking their tech to the big time. What one piece of advice would you offer them that you wish you’d known when you were starting out? 

If you’re on your own, you need to go and find a co-founder: very few businesses are investable if they’re sole-founder businesses. So you need to go and find somebody that you trust to work with, who complements your skills.

Remember too that as a startup, you are not a business. You are at best a ragtag group of people trying to find product-market fit. Recognise the difference. Do not busy yourselves with all of the expensive and distracting elements of being a proper organisation. You can do that later when you know you’ve got something. Focus on taking what you’ve got out to other people, showing it around. If you don’t cringe when you show it to somebody, you’ve left it too late – you’ve over polished it. 

Get out of the building early. Share your ideas. Prove product-market fit, and when somebody’s willing to pay you for it or actually say how big a check do I have to write for you to not take this away from me, then you know you’ve got there. Then you can start being a business.

Looking for further insight into start-ups and transforming growth through tech adoption? Email us to find out more at [email protected]

Rate this article

 

 

EXPLORE MORE

Got a question?

We would love to hear from you, feel free to get in touch with our experts to find out more!

Ask us

Interviews

Digital and technology figures share the complexities of operating in such a fast-pace industry.

Access knowledge

Subscribe

To our monthly newsletter

Sign up