Justin Lyon, Founder and CEO of Simudyne met Katia Lang, CEO, The Fintech Times to discuss the business benefits of modelling and simulation for crisis prediction and aversion.
KL: Why is it important for businesses to adopt modelling and simulation strategies for decision making?
JL:The idea of computer simulation has been around for a while now and I have been building simulations for the last 15 years. The simulations I have built range from financial markets to analysing traffic and energy flow on smart city grids. At Simudyne, our thesis has been that in the 21st Century you have to make critical decisions which can be made much more intelligently with the information that simulations provide. So I have been building simulation platforms for a range of quite interesting clients from around the world.
Simulations are how we train our intelligence, whether human and AIs. Your AI is only as good as the simulator in which it has learned its decision-making policies. Given these limitations, I realised over some time that there was an opportunity here that would allow that kind of decision making in the 21st Century. So we set out to build a next generation simulation platform.
KL: How did your career path lead you to modelling and a simulations business?
JL: I am a serial entrepreneur! I have built three companies now. For 15 years I was busy establishing a career as a creator. I started by building websites after attending university in Missouri. In 1999, I realised there was something dynamic out there happening in Silicon Valley. I joined a dot.com backed by Omnicom which we took public on Nasdaq in 2000 – that was our first taste of success. It turned out to be lucrative for all the employees i.e. we had a good opening day on our share price and everyone became liquid and sold. It was exciting. After that, I enjoyed life for a while in Vail, Colorado. Then 9/11 happened and things rapidly changed.
I went to MIT and took Mathematical Modelling on terrorism, which took me further in working with the MOD on counter-insurgency – in the fight against radical Islamists, affecting the deployment of teams in countries such as Iraq under General Brown and working alongside US homeland security. Since then my experience with simulations has led me to operating in such disparate fields and industries, developing simulations for disaster aversion and control. For 5 years, I worked with the Bank of England. I have worked with Port operations of Seattle where we actually simulated the destruction of the port. I have worked with the Centre for Disease Control and with an oil company in the Gulf of Mexico. I took on a project in Kuwait for two years where I worked on simulations of the Arab Spring and the impact of this on the balance sheet of a Kuwaiti bank. I have to say that my work and my travels have taken me to some weird and wonderful places.
KL: So what attracted you to get back into tech start-up mode?
JL: After 15 years of working with the DOD, the Bank of England and major software companies, I realised I had been using a lot of off-the-shelf ranges of software from basic, cutting edge, open-source, out of network, to niche and highly expensive. Yet coming from a startup background myself, across all of these areas I could see where the gaps were. Glaring discrepancies presented themselves when there was a requirement to build something big for the Governor of the Bank of England, a chief economist, a chief risk officer of a global bank and a Prime Minister. The challenges then lay in creating technology that needed to be different i.e. robust and scalable. That’s when the idea hit me, and I questioned “why don’t I do that”?
Four of us in partnership built a MVR (minimum viable product) and on the back of that we raised money and applied to a fintech accelerator, TechStars in London. London is where I started my company Simudyne, launching the software 1st version in 2017, where we went through a three months accelerator programme. We approached Barclays Bank to take our product on and they put us through our paces for 9 months before buying it. That trial finally paid off as both the CEO and Chief Risk Officer have publicly endorsed it. Now we have a major card processor and an Australian bank amongst many other well-known brands, as our customers.
KL: What’s the need for Simudyne expertise and simulations?
JL: I find it strange that companies may not want to talk much about using our product. Why is that you may ask? Well, it’s because it has to do with the evolution in mismanagement that no one wants to address. For instance, one only has to think about the glaringly obvious 2008 financial crises and how companies and the economy want to avert the next catastrophe. So we see simulations as putting into place preventative analysis; it’s about foresight and what could happen in the future, which up until now the market hasn’t been able to work out, for all sorts of reasons. The 2008 crisis has driven entities such as the Bank of England, US Treasury and the European Central Bank to reassess via a whole range of simulated scenarios, “how to survive”.
The world already understands about training humans such as pilots on computer simulations of flight scenarios so they can have a ready solution, in order to survive. Simudyne can build a hyperdelic simulation that captures the real world, that allows CEOs and risk management officers to explore all sorts of plausible futures, by seeing how a crisis could unfold, hence giving them a playbook of decisions they can use to choose a preferred future. Simulations can inform and benefit an organisation. However, humans can only train meaningfully for so many times and hours whereas in training AIs, this unlocks real value as they can play 100s and 1000 times a day and in so many permutations.
Kl: What is your strategy for banking and finance institutions?
JL: Banks are in the business of two main things, managing risk and making money. Banks have a lot of data to put to the test and want to use simulations to see an aggregate impact. In building a simulation, you can recreate the world in quite some detail. Currently, I am building a mortgage-model, For instance, there will be people taking out a mortgage and each householder has a different annual income. What we simulate is if there is a shock to that income and people cannot make their mortgage repayment then the bank can plan and calculate for its ROI and accounting write-offs, by studying the volatility of that income.
Our modelling approach is agent-based. Moreover, The Bank of England began collating its research on crisis prevention from lower tech levels ten years ago, but even these traditional methods such as machine-learning are not enough. As Alex Brazier, Executive Director for Financial Stability Strategy and Risk and a member of the Financial Policy Committee (FPC) has proposed in 2018, “Don’t wait, simulate.”
The Chief Economist at the Bank of England says simulations work, so I would say machine learning and simulations must work in conjunction, hand in glove. Robust predictions will be a way for policy makers and politicians to make better decisions by identifying plausible futures in light of shocks.
KL:What happens after Brexit?
JL: Brexit could be a big opportunity. There are many companies building models currently to understand the impact of Brexit. We are a tool provider, we build software and offer this to banks and hedge funds to accommodate to their own requirements. In terms of Brexit, financial institutions may look at counterparty contagion risk or a hedge fund may look at what might happen under an asset price decline, as a result of fire sales. So managing risk and making money are two use cases combining machine learning and simulations to seek Alpha. I see it this way, these are the guys that are looking to find the gold and we sell the picks and shovels for them to find that gold.
KL:So what future plans are there for Simudyne?
JL: Our company has grown from 4 to 30 people quickly. We are fully invested in serving the finance market because it’s important to the world and touches everyone’s lives. I believe, if we can manage our macro financial systems stronger, we can make meaningful changes across the board. For example, if simulations can help manage the environment and weather conditions, then we can avoid quite a few natural catastrophes. This could make strides into ensuring food security. There are so many other financial and regulatory considerations, to avert another global economic crisis.
KL:So what’s your personal goal with all this?
JL:I think my goal is to make radically better decision-making, a reality – we really need it as a species! All joking aside, I want to be able to say, “Computer what is my next course of action?” I want the computer to run millions and millions of scenarios for me to suggest my next best move. I want to take that suggestion with some assurance that it will help me move forward with my objectives.