 Thank you all very much for your attendance. I'm Alex White. I'm delighted to have been asked to chair this session, this presentation, and to welcome Mansour Hamarion, who is our guest speaker this afternoon. I think we're all familiar with the sustainable development goals. In fact, Mary Robinson spoke here last week, and we're very interested in the presentation she gave, and I think it was an opportunity for us to reflect on the, and as indeed is today's talk, to reflect on the sustainable development goals, particularly goal 7, with which we're all familiar, which commits countries to, quote, ensure access to affordable, reliable, sustainable and modern energy for all, and you'll be familiar with the fact also that the UN in its 2017 report, or its assessment, pointed out that in 2014, 14.7%, just under 15% of the global population, has no access to electricity. Well over a billion people, predominantly rural dwellers, have no such access, and half of these people live in Sub-Saharan Africa. So that more than 3 billion people, mainly in Asia and in Sub-Saharan Africa, still cook without clean fuels or efficient technologies. The World Bank has identified that emerging and innovative energy service delivery models offer unprecedented opportunities for private sector driven off-grid electrification and for accelerating universal electricity access. So the urgent necessity for progress and for increased and enhanced progress is clear. The close relationship between access to clean electricity, on the one hand, and material, basic material well-being on the other is, I think, clear. The opportunities that exist are manifest, especially given the extraordinary evolution that we've seen of data technology. So who's going to lead this? Who's going to do the work? Will it be governments? Will it be private sector? Will it be people in the academic institutions developing new products, new ideas, and bringing them to market? Who and how will we take up the enormous opportunities and challenges that are there? Bee Box was founded in 2010 by three students at Imperial College London. The organisation today deploys off-grid solar appliances combined with its smart solar platform to bring machine learning and customer experience optimization throughout rural Africa. They have over 100,000 systems deployed so far and 450 staff across five offices in China, the UK, and East Africa. Bee Box is one of the words, literally one of the words, leading off-grid solar companies. We have with us today co-founder and chief executive of Bee Box. We're delighted to welcome Mansur. I'm not going to go through his extraordinary biog, which is incredibly impressive. He's born in Pakistan, raised in Sweden, I think, studied electrical engineering at Imperial College London. He was the founder and leader of a student charity E-Quinox, which brought electricity to six villages in Rwanda. That was just when he was still a student. Following university, Mansur worked as a manager for Rose Royce, civil aviation side, and on corporate cost reduction in the Asian region service and the overhaul business. An incredible stellar biog, CV, already for a very young man. We're, I think, all very much looking forward to your presentation this afternoon. Will you welcome, please, Mansur? Thank you so much for the kind introduction and for setting the tone as well for this presentation and for the invitation here. Great to be in Dublin after 10 years. Very different context. Last time I came here, I was a student. So a lot more professional this time around. So what I'm trying to tell you guys about is not just what we do because that's going to turn, that's going to become very boring. What I want to really tell about is the big opportunity that we are facing as well. The problem that we have, the opportunity that we have, and how to look at energy in a slightly different way. Obviously, I'm going to give some examples of the sort of customers that we have, but I think the themes that I'm trying to tell you about has a relevance in the overall energy economics of the future and data plays a very big role. So let me tell you first about the problem that we picked. The big problem that we picked was that over a billion people around the world have zero electricity, none whatsoever. Further, two billion people have unreliable electricity. So they can't trust the supply of the electricity that they have. And that's something that Bebox, when we started with my two co-founders, Chris and Laurent, we found that to be an unacceptable reality, that in 2010, 2011, when we got serious about it, that how can half the world be digitally excluded and one part of the world having smartphones and all sorts of modernities, another half of the world doesn't even have an opportunity to engage with that. That was the social or the moral conviction for setting up the company. And what we focused on really was why don't people have electricity? Is it policy? Is it technology? Is it what is it? But fundamental to that, what we really wanted to do for us was to understand the customer. And that approach of trying to understand the customer actually resulted in that data-driven business model. I'd also like to be honest today is that when we started the business, we didn't see the data as the core part of our business. The data was just a means of trying to understand the customer. But it transformed itself to become the core part of our business. So with that, let me try to explain the type of customers that we are talking about. These are customers, 600 million of them in sub-Saharan Africa, that already spend money on energy. But they spend it on really inefficient, really unhealthy, really polluting sort of energy. They spend it on kerosene, diesel, firewood, disposable batteries for their radios, disposable batteries for their shavers. They spend it on a phone charging. The mobile phone part was one of those things that actually grew an explosive way over the last decade. Actually, there's over a billion SIM cards in Africa, but there's only 600 million people without electricity. Actually, majority of the populations have some sort of remote access to electricity where they can send their phone for charging or buy a disposable battery. But the fact is that they don't have it in their own homes. They end up spending a lot of money on trying to get the basic charge. Just in terms of math, what does that mean? I would say that the 95% confidence interval on this data that we have is that they spend somewhere between $5 and $20 a month on kerosene, candle, battery and phone charging. That's the sort of consumption profile. And then we have some more extremes as well. But the most terrifying part is that the majority of customers are spending more than 60% of the truly disposable income on energy. So these are people that are living in extreme energy poverty. The majority of the world in the developing world, their main expenditure is energy. And that's a terrifying fact, but also an opportunity. So, if I click this through, all right, there you go. So the other thing we wanted to do was we just don't want to solve electricity. We want to be able to give people in the rural world access to a modern lifestyle. So what we realized quite early on, and that was the foundations of the business, but actually electricity is a Trojan horse towards many other value-added services. It's the start of having a TV, a fridge, a business, employment, jobs, et cetera, a shaver. There's tons of things happening. And then another thing happened over the last 10 years, or two things happened over the last 10 years that made our business model possible. One was growth of mobile money. I don't know if people are familiar with mobile money, but just to explain what it is, it's a very African phenomenon. It started in Kenya. It was a company in Kenya called Safaricom, the big mobile operator there, that started realizing that customers were sending their air credit to each other, not to top up another person's phone or the relative's phone, but as a means of payment. And now that's been branded as EMPESA, and now nearly a third of the Kenyan economy flows through EMPESA. So mobile money became huge, and this has become an African phenomenon. The second thing was the growth of phones. Cell towers and cell signal became quite accessible in many parts of Africa. So our ability to actually have an IoT device, remote monitoring, remote control suddenly became possible too. So that means that we actually were able to not only offer appliances that are becoming more and more efficient as we all know, your TV, your light bulb, et cetera, but we're able to actually combine that with remote monitoring, remote control, and IoT solutions to switch them on and switch them on to be able to actually create an economic proposition for them to start with electricity, then to be able to actually provide them a TV, a small fridge, a fan, whatever they might need. And that's a sort of value proposition that wasn't possible to do a few years ago. And I'm going to go through a few distinct examples about that. But what does this all result in? So this is the future that we imagine. We actually have installed now 160,000 households with solar power, so individual solar, battery, and a range of appliances. Each of the households is remotely monitored and remotely controlled. So we can see exactly what the consumption profile is when they pay us, we switch them on, when they don't pay us, we switch them off. It becomes an easy financing mechanism for that. So we estimate around each household, each system impacts around eight people. So actually tonight around a million people have access to electricity, which is the cool part. And every day we are installing 200 to 300 new homes per day. So we've been accelerating quite a bit. The byproduct of this is the data. So this was the way to actually offer a customer, hey, take this product, we will install it for you when you pay, you're going to switch off, when you don't pay, it's going to switch off. The resulting fact of this is that we have these customers who are not just energy poor, but no one knows them. Their average is on the World Bank data, they are unknown to the government, and suddenly we have this super data-rich business model where we're seeing all sorts of data for the first time to actually understand the customer and to be able to see what their ambitions are beyond and above a KYC or an ID document. And that was the big opportunity that we personally did not see when we started a business many years ago, which was like who is our customer, why don't they have electricity, let's build tools for us to be able to gather this data, and let's build tools to be able to offer value proposition to the customer where you can understand. But the resulting fact of that was we actually collect, which might not mean much to you guys, but it's five billion data points per day from the most underserved people on the planet. And I'm going to show you some of that data. So this is consumption data. What was the consumption? Is it behavioral data? How does the load profile look like? Is credit data? What does payment profiles look like? It's also staff management data. So all of this data gets transformed into actions. And the resulting future is this virtual utility where people have their own self-generation, they have their consumption, but they still have a relationship with the central body that's managing them, that's financing them, handling their problem. And I actually think this is the future that the entire energy of the world will move towards, where the cost of cables will outweigh the cost of batteries, the cost of cell generation will become independent. So the interesting thing is in the same way that mobile payments started first in Africa, and now we can do it here. And that's the same sort of thing that LeapFrog that Africa is doing with the modern generation will happen in the western, in the whole world as well. Obviously it's going to be a translation process, but the big themes are going to take place. But the exciting part is we already can demonstrate some of the characteristics of this future. So what do we actually do from a product perspective? Just to let everyone visualize, we install a solar panel, whatever the right size is. We install a battery, whatever the right size is. So that means that they have electricity both during daytime and nighttime. And we then install a range of appliances starting from light bulbs to TVs, to radios, to shavers, and we finance those appliances too. So if you just go and provide electricity to people who've never had it, they're not going to have the appliances to run it. They're not going to have the money to be able to go and buy a TV. So it's important that we play a role in the appliance side as well. And if appliance breaks, they have nowhere to go and service it. So that's why we need to also provide the servicing to the appliance side as well. So what we do is actually, in a sense, the modern, the European utility model is that the appliance is a separate story from the generation and a customer is dealing with a ton of different providers. In this sort of virtual utility, you are actually responsible for end user experience. And that, I think, is not only an African phenomena, but something that's going to, that behind the major sort of activities is going to happen more and more around the world. Where does our data side fit in? So we actually now have crossed, it was cool when we said we have crossed a hundred employees. That was really exciting. A hundred. We felt mature. Now every time I see the data how many employees I have, part of me I'm like, oh no, it's too many people. More people to manage. We just crossed 600 people now in the company. And we have 50 people just dedicated towards software and data science, all based in London. What we're doing is basically a big data play. We have each of our devices, so 160,000 devices that are sending us data every four hours of what's happening. We're seeing all this payment trends happening. So it's in large amounts of data where we can actually generate insights from. So let me show you some of them. So this is an actual customer, a not named customer with the GDPR rules, things are even more strict. So what we, on the green side below, this is the days he, his major was positive. And the moment it turns red, does it have a laser thing? Well it does. All right, great. So here you can see it was on until the third of September. His system got turned off. He made an immediate payment and it got turned on again for a few days, right? So he was positive. Then actually we sent him a warning signal again so that the orange side is like, we are going to turn you off if you don't pay us. Now in the beginning you can see it's a very small one and he reacted immediately and he paid. Here it is a very long and now until at least end of October he hadn't paid. It's an example. The key question that we're trying to answer in the business is what happened here? Could it have predicted this? All right. And what exactly could we do to be able to extend his temporary loss of income? Is this temporary loss of income? Is this sustained loss of income? And what is a sort of business model one can develop to be able to do that? Maybe we financed him too many, like he might have like one TV and too many light bulbs. Maybe we should take back one TV and a light bulb so he actually can get back into a basic need. He might have too many things, right? So that's sort of like capacity to pay issue in an individualized basis we're becoming really good at. We can look at any customer we have in our portfolio or any new potential customer and to a fear to statistically relevant accuracy tell anyone how likely are they able to pay and how much are they likely to pay. And those are two very important questions regardless of the underlying energy economics that unlocks financing, unlocks our scalability, allows us to manage risk. And this is the way it works is that if people familiar with machine learning techniques, it's a big black box. Don't really know how it works. If we throw in all this data into it, runs a few statistical models, every day gets trained it becomes better and better and it retrains based on historical data. So the more that we collect, the more different types of customers that we get, we become better at doing this. And with some of the modern techniques coming on board, it becomes faster to do this learning as well. The second thing is a connectivity. So this is an example from our dashboard. There we go. So this is in Rwanda. You can see the concentration of customers that we get. So we open up a shop and we start to like install each customer. And each customer has a GPS location on the product. So we know exactly where the system is, what's happening and each system sends back data. The data is of this sort of profile which we can see exactly with the current, the voltage on several different components. What's really exciting about this and this is something that the utilities of the future, regardless of the geography, have to prepare themselves for. Which is that something again, we learned from our mistakes, not proactive planning. I want to be very clear about this. A lot of this stuff is not because we sat down seven years ago, hey, this is how it's going to look like. And this we know all the variables and we went and did it. A lot of this from mistakes that we learned. One of the big mistakes we learned is that in a distributed generation world, all your problems are distributed, right? You can't send your best engineer to your centralized power plant to fix it, right? And you can't send a person with all possible tools to fix it. All possible parts to fix it. You have to send the person with the right tool, with the right part of the right location in rural Africa to manage it. So how do you solve that question? It took us years to solve this one. Which was that we have to turn all of the problems proactive. We have to live in the future. So every single system that we have has a twin in the cloud, if you can imagine that. Where we look at this current technical performance, the state of its battery, the state of its electronic, and we simulate things into the future, and we look at what's going to break and when is it going to break, right? So that means each system, we can see exactly what the risk is on each system. So that means at a national level, we can do appropriate spare-level, spare-part provisioning. Whether it's like components of a resistor or a fuse and things like that, or a battery or a solar panel. So the country itself has enough spare parts. Then when things fail, the system can diagnose itself, 70% of the time, to tell us what actually failed and what part needs to be replaced. Or if it's a customer-induced problem. One of the earlier problems that we had from our customers who liked our system very much was they used to throw a bucket of water over it on a daily basis. That's how you wash things, right? And our systems was not designed for a bucket of water. Second issue that on a daily basis, sometimes it's okay, but like over a course of actual, a course of period, things happen. Other issue we didn't really take into account was mice and rats. I like to think there's no mouse in my house or rats in my house. In London, apparently, we're not 10 meters away from any mouse or rat so far hasn't seen any. But in Africa, in rural parts, wildlife is much more closer than we expect. Cables that become slightly warm are attractive to creatures of that sort. So those are sort of things that we started learning like, hey, something we can design out and become better at or teach people about how to do. Other things we have to become more predictable. So, and this is basically the translation process. So, hey, we have one system, all the current voltage, et cetera. Using the modern techniques that we have, we'll be able to say that, hey, this system actually needs a battery to be replaced. Ask the customer if he has any problem. So this is as proactive as you possibly can get. The customer does not even know he has a problem. We are going to proactively call him, tell him he has a problem, and I'm going to organise a replacement date four or five months into the future. So this is a sort of level of things that are possible. So that means that suddenly we can manage a distributed workforce. Our super intelligent engineer is no longer a human. It's this black box that no one really knows how it works that actually can look at all systems at all times and all seeing algorithm that actually can tell us what the system is going to go wrong. And then we have another black box that tells us which order we should do things. So the scheduling for when you have, we have 600 full-time staff, over 2,000 Uber drivers, in the lack of a better word. They have our apps so they can sell, sign up customers, they can install customers, they get paid on a per action basis. So we basically do all the scheduling for them using the system as well. So this was the technical backside of it. And it was not something that, if I knew that we had to do this in 2010 when we started the business, I would not have started it. So ignorance often is a bliss for sure. I had known through about half of the things that I'm telling you that we've done right now when we started, which is, which I always tell my investors is a good thing. So in terms of, in terms of, this is one of those things which I find really, really exciting. So I'm a bit of an emotional character. So in the beginning when a customer used to pay, I could see his phone number, I could use his user profile and I nearly wanted to call the customer and thank him for using us, right? Now we have a lot more, so it's a bit more difficult to do that. But this is one day's consumption profile on a group of customers that we know live below the poverty line. So this is a group of customers in East Africa. We define the poverty line as $1.6 per day. There's slightly different definitions of it, but poor people, regardless of how you're going to put it. These are people that are definitely below $2 a day, but we define as $1.6. And it's an aggregated customer, so they have no TV. They won't be able to qualify for TV. They're too poor. But I just wanted to show how, one of the hopes I had with this data is that we come with this magical insight that the behaviors of the rural poor is of a certain sort. The really boring fact is they're exactly like us. Like there's very little differences, if any. You can see that around six o'clock people start using the system the most, right? Lights, et cetera. Most people go to bed around 10 o'clock. They wake up much earlier than I do. Some of them are awake around four o'clock period. And then you can look at how the profile actually looks like. And that means actually what we're trying to design for, and I think that's been one of the big, big, big problems in development is that we often put the sort of customers that we've had as beneficiaries, as not customers, that their needs are somehow different than the needs of ourselves, us and them thing. The data has shown us, it's exactly like me. Like, I don't wake up at four, admittedly, but otherwise it's very, very similar, right? And that makes it quite exciting, even more exciting because the sort of solutions that we're developing, the sort of things we're developing, actually has an application far broader than the geographies and the demographics that we're targeting. And then I think that's a very interesting point for us internally. The other thing is the role of data and the responsibility of data, right? So this is one of the things that we've been talking a lot about. So this is a live view of each agent. So if they have a phone, the phone which they sign up customers, do the credit check, et cetera, we can see exactly where they are, how they move, what's happening. So actually that means that we can break down a country into good parts, bad parts, more difficult parts, right? Where we want to find customers, what's the right spot. The best current sales location in an entire business is a university in Kisumus cafeteria. That's where we sell the more systems, right? And we found that not because we knew the cafeteria in Kisumus university is great, is because we have agents clocking in sales there. The student group is a big group that we sell to, right? But the fact that we have all of this information, one is great, but how do we make that actionable? So what each of the agents can see as we process all of this data is exactly what's the next thing you should be doing. So that means that an agent has a to-do list. So we turn this three or four or five billion data points that we have to around 20,000 actions, actual actions per day, and we distribute that to around 2,000 people. You do this, you do that, you do this, you do that, all right? And we're becoming also about to release a feature that we even can do when to do it and where to do it, right? Now we say like sell or upgrade, go to this customer to upgrade, for example. But now we can tell him to go to this location, do this type of thing. But that means that we have a lot of data on people and how to be responsible about this, right? And we use gamification techniques and we see as people know if they rank the highest in the country today, highest in the region today, et cetera. And that's one of the things that we've been battling a lot on. Just from a regulation perspective with GDPR, et cetera, but how do we actually collect data that's meaningful to actions and to development, right? And how to be able to actually hide some of this from our need to know basis within the company as well. The biggest challenge that we have, and I tell this, we're back by some of the larger investors in the world. We have money from the Silicon Valley, from the likes of Costula Ventures, which is the founder of Sun Microsystems. We have big investors like NG, the French utility giant. We work closely with Orange Telecom in Africa and EDF. So we have big sort of capital pool that we invest and we finance our customers at a large scale, obviously. So the good news is that we don't need trillions of dollars to be able to electrify the developing world and transition into a clean future, but we still need billions of dollars to do so, right? And that's maybe the more sad news, right? And that means that we need to find, although we have around a million people tonight that have access to electricity, there's still billions that don't or have unreliable. So it's partial success, a grand pilot that we have demonstrated so far. But for us to be meaningful, important and big, we can't do it alone firstly, right? We need many, many parties to get involved, really like transition ourselves to the new future, but we also need to be able to find mechanisms by which we can actually fund countries, fund people, fund places that have not been common to do. And this is where I think the role of the data plays the biggest role, the perceived risk and the real risk, the perceived need of the customer and the real need of the customer, the perceived opportunity and the real opportunity. And I think that data for me has the potential to bridge that, to be able to actually result in not like preconceived judgmental view of investment or where capital should go, but actually a data different decision. So with that, I'll end it there, but hopefully you guys have any more stuff you can ask me questions, I'm sure, but hopefully they give you a view about what we do, how we do it and how it could be relevant beyond the shores where we work today. Thank you so much.