AI for Good Innovation Factory – LIVE Pitching Session
10:31AM Sep 7, 2020
afternoon Good evening and welcome to the fifth startups pitching session are they after with innovation factory yet see you then National communication union. And I have the privilege of facilitating this session today. It was the United Nations supplies agency for icps. And we are also organizing the AI for Good Global Summit Foundation 36 United Nations agencies, ACM and co convened by both of the summit applications of artificial intelligence that can advance the Sustainable Development Goals and to scale it up for a global impact. The inevitable innovation factory is to build and bring together an inclusive and diverse community of entrepreneurs, startups scaleups idea innovators are using AI to achieve their one or more. And I'm very pleased to note that the goal for startups is open on our website it is it empty and anyone have a solution that can use AI to achieve one or more sustainable Developing goals and meet them and life coaching sessions like this one. During the live pitching sessions, we're going to have experts on the base produce model, scalability, potential social impact and social dilemma start dose outcomes of the 2020 AI for Good Global Summit grand finale of the grand finale is going to be on the 19th of November this year. And before we start with introducing our mentors judges and say the housekeeping rules for this session, I would like to give a quick exercise to all our attendees and panelists today. So can you please use the chat function and let us know where you connected from which city or country and please make sure that you market to all panelists and attendees so we can all see and I will start with myself so I would right here
the casino is coming A lot of people are saying so we have people come from Paris, from Madrid from Munich. Or from the US from USA. That's amazing. So a lot of diversity I can see in the room. Yeah, Kenya, New Delhi, Dubai. Great. So the housekeeping rules for today, we're going to have five startups coming from different countries, and have four judges for experts that this has been we have shared with the judges this the scorecard. We can't even wait the startups using this one. And we'll have four minutes for each startup to pitch their idea and solution. And also we'll have a q&a afterwards between the judges and the startups after each pitch can be asked as well just to actually send them send us their score using the score cards. And I think I can move forward now with reducing our mentors for today. Our first just name is Josh short. Josh is an international expert in startups such as Development Partnership, the link and global marketing, particularly The ICT, and he's now director of the business development of Korea. So before, I also have to mention that previously, he was an IT colleague, and he will work actually for it. Hi, Josh, how are you?
I'm very good. It's very nice to see again, you
versus you, and really have the pleasure to have you today as an expert to check the startups you can have. And you work you work, I mean, with a lot of ICT startups, a lot of startups and also it both so much innovation initiatives and stuff like that. So my question would be, based on your experience, what are the main challenges for for iced tea or our startups? from your experience that can actually kind of limit their scalability?
Yeah, this is really a good question. I mean, I probably out better have focused on AI related startup because essentially, it's about an AI for Good. So but if you think about ideal picture of AI industry I mean, we can see four major elements of fully sample computing infrastructure, and algorithm and data intelligence, which means that, let's say, if you really wanted to build a really nice, you know, ai business today, you need a really super power computing infrastructure, which can deal with a huge data set, which can be analyzed by a good algorithm, developed by a talented engineer, basically, that's the really ideal picture, right? So, you know, computing infrastructure and an algorithm more like a hardware element, which is relatively easy to get compared with data talent, which is more like a software element. So especially for the startup, you know, this is this really, really challenging for the startup to get really qualified high quality data, because in a lot of data is really gained by the the big tech companies and then big companies, you know, the big organizations who are not always you're willing to share the data within other small startups, because there is really Important fuel for the business right? So basically the startup has to find that their own really novel way to get really high quality data by for example, maybe you know launching their own MVP minimum viable product so that they can actually test with the customer who may be you know, if they have a really good relationship with your data owner, probably you can build a partnership or something but anyway, you know, so to getting the data is is really crucially important but i mean it's it's one of the big challenges for the small startups wants to actually use the data and then the category a meaningful data good enough to train their AI to death probably to one area of challenging and secondly, basically talent, you know, actually was given the really high speed of development, the AI business industry. Now always start off as a really heavily having, you know, are struggling to find a really good talented engineers and developers and in filling for date, luckily get them in a day are always necessities to To get really high or higher higher compensation package by the big companies, and then this is one of the big challenges and also and the current education system is not well really optimized to really nurture AI talent these days. So basically, the government and education an expert has to has to really find the best way to nurture AI talents, what are the really, you know, the the missing gaps between the real need by startups and the actual, you know, the the students and actually creating from the University with a really relevant skill for them for the AI industry. So, I think this, you know, two things in determining, you know, talents aren't really my biggest challenge for the startups, especially who are considering the use of AI for their business.
Yeah, very good point. Especially it's it's interesting to talk about data and actually give data and giving really high quality and relevant data and also, that you're finding the right time for startups. And this makes me ask these these challenges are two specific To ICT or HIV startups for it's also for other startups in general are actually suffering from different challenges between AI startups st and others.
Well, I mean, um, well actually what I started from the perspective you know, with my answers, but basically I put, I strongly believe that all the startups should be equipped with AI arms in the near future anyway. So basically, because basically AI is really about you know, helping you make the best decision, you know, and to find the Nobel you know, insight and solutions by developing artificial intelligence to help you so, whether you are, you know, running any what kind of really business you are doing and basically, the AI is really essential part for you. So, I think this challenging is not really something only for the, the the AI really based truly in a tech step or rather, whether you do business on your platform or digital economy, re commerce and whatsoever data Talent are really, really important this set for you to get. And I think this is why I think the these are the really challenging photo, mostly all startups.
Okay, great. Thank you for thinking about Josh and looking forward to the question the pitching. And I'll be making that. Thank you. And viewing for our next expert is actually john pushya. So Dr. Jonah shear has more than 25 years of experience in writing global roles, including executive leadership, business development, sales marketing, is currently works at Dell Technologies as head of run the business intelligence transformation. JOHN, thank you very much for being with me.
Thank you very much army for inviting me in. It's a pleasure to be here to be a judge on your panel. I think you know, in relation to cognitive psychology, when do you look at cognitive psychology in relation to AI? There is a perception here so I'll give you a story what we have tried to do over 24 years, right, so we've done a lot of research in compression technologies, number one Lip Sync technology number two facial expression technology number three, and we combine this into virtual characters. And we've tested this out now in various so this has been going on for the last 20 odd years. And we we've combined from the from the elements of cognitive of perception of stimuli, of problem solving, and how these characters can actually develop would actually would sit with children and work with them in HR issues, and interact with them now in the last couple of years, the AI has evolved so much that we've come to a per diem in some degree in our in a way of our research. And we are putting a lot of emphasis on the perception element, the emotion element of the virtual character. So that's the psychology cognitive power tools. And then in relation to the AI side of it is it is an analysis of the data that we're actually getting, and how we can persona characters can perceive the environment environment also work. In an element that they can actually be like a virtual twin. that's becoming a reality for us.
Wow. Okay, so like, have you started between psychology? I mean, I think you have a PhD in business. So yeah,
I'm a climber. I'm a psychologist. I'm also a technologist.
Well, that's a great read. Actually, we haven't one of the startups today I think the work on ADHD. I leave it to them to explain so yes, it's really interesting to see this kind of thing. Maybe a follow up question quickly. This is a business opportunity for startups and innovation in this specific domain that overlap if you
will of course i mean if you look at if you look at entertainment, right entertainment, as they call it, I mean, hollywood are I mean Pixar, for example, are pulling a lot of startups right now. Looking at the under technologies because you need a is very important in relation to computer compression and everything else technologies and processing power and all that If you can create certain algorithms, especially in deep learning nor learning neural networks, it'd be fantastic. I mean, that's my advice, look into neural networks, because that's an area where it's actually booming right now. When you get into video gaming, look at the emotional the behavior and patterns of these characters that are creating. They're getting more still get more stimulation of that. So that's something that's actually grown. No, unfortunately, because we have this COVID-19 situation for us, you know, more people that are online, they're watching more games, they're playing, they're entertaining. There's more startups actually come more into us in psychology, cognitive psychology, perception, memory and problem solving, etc, into their game and perception so that there was definitely a boom there on this was also from a healthcare,
sell healthcare. I mean, like nurses prognosis, every fashion was amazing. Its enormous. Thanks, john. Great to have you today. And look pitching. And well, we're waiting for questions then. And thank you. Have you moved forward with our next judging expert, and she's tenure wise, so tennis wise, and he is the founder of IoT tribe, Internet of Things tribe, and he has a strong track record of almost 20 years in conception launch delivery programs, support startups. She helps clients define theaters, and create ecosystems for innovation across UK and Europe, Kenya. Thank you very much for being with us today.
As a pleasure, thank you so much for inviting me.
They have a lot of experience in building and delivering proven that actually supports startups right. And the AI Summit innovation factory is all about the AI startups that can actually achieve a state of the world. So from your perspective, what are the factors that defines for you innovative, scalable, promising AI startups? Well,
that's of course the $6 million question or $200 million question that every startup and don't buy wants to know the answer to. But I would say that from from our experience with an RA IoT tribe, there are three very specific things that we've seen that are necessary to fuel that energy and that commitment to the long, long journey that is the startup growth. The first of call is, of course, is the ambition of the founders. So the founders, you know, you can have founders that come from r&d, you can come founders who have no experience in building technology or technological products, but really clear on their markets and the problem they want to address. But regardless of what their background is, they do have to come into their ventures with a whole load of passion, energy and belief in themselves because they are going to encounter a lot of problems, problems that relate to the product or ability to communicate what they do financial problems, and if they don't have that self belief, then the journey is going to be that much more difficult. The second element, and I do say this, the second element actually, is there has to be a market for what you do. So you can have this wonderful technology that is the best at whatever it is in the world. But if you don't have a market for that technology, if you don't know what that route to market looks like, again, it's probably going to be something that remains interesting for, for study, or for a pattern, but isn't really going to be the basis of a good startup. And then the third thing, and I would say this is specific to AI companies and the ones that use deep tech is that you have to build the future. Now what they mean there? Well, Josh mentioned the triumvirate of almost any technology startup. But first of all you have to have, if you have hardware, well hardware and if you don't have hardware, normally you have to know how you're going to plug into legacy systems and relate to how Second, you have to have data. If you're an AI startup, there's there is no way without data. And thirdly, you have to have talent. Well, I would add two other things to that. You have to add explainability and ethics into how you build your company. Okay? So explainable AI is not just something that is for lefties or for people who have a social conscience, it is something that fundamentally is going to determine whether, for example, a large corporation can integrate your AI into their business processes, because from a compliance perspective, it is quite possible that years down the line, if something happens, either a law court or a licensing authority or any type of authority with which you with whom you have to deal is going to ask you, you know, why have you done this or why has your product ended up doing this? And if you don't have the answer to that, then that's going to be a problem and ethics Of course, huge, huge issue with with AI, making sure that whatever it is you're building, you have unbiased data. And if you have bias data to begin with, and you very, very quickly understand where that bias is, and you take steps to remedy and to eliminate as much of that bias as you possibly can. And of course, you know, just continuing in that long, long term process.
No great points. I like that the first thing you said is the ambition and actually the self belief of the the team or the cofounders. Great. So maybe, I mean, looks like a very tough criteria to define a successful stadium starter for you. I mean, it's a follow up question. Do you have any examples in your mind?
Music Festival, ai startups?
Yeah, absolutely. I mean, we're very lucky and I teach five ways. We have a tribe of close to 300 startups now. Operating quasi seven things AI need. Now we're starting to see some in quantum computing. But let me just pick out two of those. One of them is called punk cue. And they derive economic intelligence from unstructured and alternative data sources. And what they do is they essentially enable companies and organizations to have foresight of things before the rest of their market. So they can predict what US GDP might be, or friendship, inflation. And they can also really pinpoint from a sectoral point of view what's happening within certain geographical markets. So for example, they might be able to tell a company that wants to set up operations in China or is planning what resource they want to place in China, what the industrial production is going to be in six and 12 months time. And of course, that product is used really, really widely with investment strategists. So looking at what They tactically base their assets. And you know what sectors are likely to do well, and they use AI. They're an amazing company, an amazing team. And then in a completely different sector, for example, we have metal slab, they have a product called Kelvin, which helps engineers in factoring, sorry, in large factories and manufacturing processes to achieve consistent product quality. And that essentially has a huge impact on the environment. Because of course, you reduce scrap and rework your produce you improve production capacity, you accelerate training of operators because they don't have to have all of these multiple things to worry about. At the same time. They have all of the key data on a dashboard. And in addition to that they offer API so that other systems can be built on top of what they do as well
as super interesting. So one, predictive AI kind of demand. It wasn't originally I think, well sci fi to actually leave in the chat some of the links of these short clips as well. We'll do.
Move on. May I just say that we've got a couple of acceleration programs open so everyone anyone's Welcome to apply to those as well. Very happy to help. Thank you.
Okay. Sure. Thank you very much, Daniel. We'll move on with our last and definitely not least, excellent yesterday was your van sturgeon which so event Thank you very much. We must leave and actually is a tech entrepreneur with many years of experience in industry, founder of Wonderland AI Summit and founder at Serbian ICT and one of the leads at CGI ran Thank you.
Thank you for the invitation admin
nerds it's our pleasure and Iran you're the founder of the Serbian AI said yet I'm really excited to know what is the Serbian website?
Yeah. Serbian as societies. Our new organization we're actually we sell that our research research community is not made Be so well organized in a sense where we believe that if we find a ways how to join forces in a totally new way, we will be able also to be a significant player on the global market, regarding different kinds of research topics, but also to educate our ecosystem, our companies, our citizens around AI covering different kinds of topics, from really deep tech research in the parts of the like ethics, psychology and other stuff that are really important. If you're talking about about AI. We saw that we had really top like researchers are like people from our country that are all over the world from Cambridge, deep mind, MIT, Stanford, Berkeley, that they're really eager to help like and boost our ecosystem here. And we try to connect the dots to try to connect them in a new ecosystem and we succeeded so We have more than 120 members who are with us there plus 100 more that are now in the process. This year situation with the call is it's a pretty tricky for for all of us I believe. But now actually we are starting with different kinds of initiatives that will cover topics from the business ground. And we are we are looking for the ways how to educate our intrapreneurs from the really top top level okay how to prepare your company for AI how to have KPIs that will how to be important for your return of investment. The second element is in the order department is related to the research we are actually this topic AI for Good or AI for social good is something where we would like to bring your contribution to the world AI. We are now we are preparing some projects. We are our first project will be related to the kids with mental disabilities. We are we want to help them to have a better assistive tech analogy that will help them to house a habit, more proper communication, etc. Of course, we are preparing other departments that will actually deal with some attic ethics questions, and some others. But yeah, we have lots of fans, while a lot of really interesting people that are really eager to be in a position to make changes, and for us this is really motivating to go further with our story.
Just to note, maybe before we start with the pitching a very quick follow up question for you, since you work a lot with ecosystem or national level of this. What are we the main pillars do you think are necessary for successful AI existence just in less than 30 seconds?
Yeah. First of all, if you're talking about this region, actually I started really lean with like initiatives in our city but then we saw that we need to also join forces with our region. Dr. wrenn to try to connect the dots in the sense, okay, some really quality people are around you, we need to open the dialogues around AI, what are the challenges that are they they are they are facing with, if they are from the side of the talents, data sources or business models, we just need to open open dialogue. With that, of course, we also need to find a way how to motivate people to share their own ideas because for example, from a couple of our events, we had the situation when one guy or key actually said, Okay, I'm looking for a co founder, I'm a technical guy, but I don't have like a business, any kind of skill set for that. And then another guy actually in the public said, Okay, I'm looking for a technical guy to support some of my ideas, and they actually succeeded to create their own startup, to go further etc. and I strongly believe that those kinds of ways of networking We are we will be able to share the ideas to meet and to see okay, what are the achievements in our region or even wider? That is really, really, really strong strong point there of course these types of events like summits, conferences, like meetups, etc. I'm like in person, like live events for me are the best. But yeah, it will be really challenging to see what's going to happen like in next month with these kind of online sessions there. But definitely we need to find a way how to have a better dialogue in the ecosystem.
No, we're trying to provide this platform even virtually with all these kind of elements to build a bit nibbling. Thanks again you very much for being with us. And thanks very much to our mentors and judges. And I think without further ado, to move forward with our pitches. So again, as I mentioned, we have five startups today we're going to have a greenhouse solution from the UK psychiatrist from Spain of your diagnosis from the US and TV from Germany. Hi, bream. How are you? Hello. Hello.
Thanks for the job, the present.
Well, thank you very much for being here today. We're really good to see you. As you know, you have the four minutes time these slides Get ready. And then afterwards Yeah. And I think I would say the smoker. Yeah.
Okay, I'm just confirming that you can see the slides. Yes. Okay, thank you very much. Thank you for the time to present. We are delighted to present an overview of our technology which supports the United Nations Sustainable Development Goals. Our solutions take into account ways to replace carbon and methane in the energy ecosystem. Which will have an impact on millions of lives, especially in cities. Imagine a world where molecule fuels and chemicals come from renewable energy sources, further decarbonizing the non electricity aspects of the global energy system. Well imagine no more green hydrogen made from renewable energy can have a significant effect on decarbonizing the global energy system. Advances in technology have led to significant innovations that improve the efficiencies of electrolyzers. thereby reducing costs but the key challenge remains finding renewable energy for a sustained duration. And linking green hydrogen to markets and green hydrogen solutions we combine big data and artificial intelligence with innovative technologies to develop lower than average cost projects that produce hydrogen at competitive prices linked to demand. We are developing a green hydrogen neural network a global neural network link clean, renewable electricity, and markets to electrolysis seamlessly across many systems and devices. So what does this mean for the Sustainable Development Goals, we are creating an efficient system powered by renewable electricity and using hydrogen to produce molecule fuels and chemicals, displacing hydrocarbons while at the same time improving energy efficiency and system efficiency. The hydrogen ecosystem is undergoing rapid development, and will over time have the capability to modernize and transform the energy ecosystem. From all the green hydrogen apps which we call them available to address the energy challenge. We can choose those that can address that can address certainty energy challenges, or backed up by data. Is that indicated before the key challenge is securing low cost renewable electricity for the longest hours and linking production to markets? This is where our solutions using artificial intelligence deployed to a smart, integrated electricity and heating system and energy market can solve the problem. We process our data on the AWS platform, and are part of the AWS activate program. We also use as your for machine learning. We have a network of AI and ml specialists working on various projects, and have advisors and engineering specialists that we deploy to develop our pilot projects.
Thank you very much.
That Okay, yeah, perfect can be a video as well so we can see you and maybe I suppose or experts if any of you have any questions to be asking. JOHN here, please go ahead.
So Amy, how are you? Very good presentation. I do have some questions for you in return To your business model, okay. Yeah. Can you define how two questions How is your business model and how we always add compared to your competitors and is there a large competitive base in your virtual Americans?
They trick questions
okay. So, so, the first the first is what is the business model? So, the business model is we develop hydrogen projects and we produce green hydrogen and we would be at lower cost. Then the current pilot projects all projects in the world basically a pilot projects because of the nature of the industry. Okay, and at present, the largest operating electrolyzer is 10 megawatts in Europe. But there are lots of announcements that have been made recently. So we would also be developing a pilot project okay on the on the technology side We would be AI as a service. So once we use a pilot project to prove our capabilities in terms of using the neural network, we can then sell that as a service. So long term, we have two revenue sources. One is we'd be selling hydrogen, at a certain price per kilo gram. Initially, we're needing carbon incentives. And then we would also be selling to other project developers not exclusively to our own AI as a service. And the reason why we have to do pilot projects is to prove our technology.
Okay, next question for me is that I've seen that you have blockchain Are you using smart contracts?
At this stage, the industry is not set up. So on the renewable energy side, things are moving very fast. Okay. But you have variable ppas are power purchase agreements, but they're not there yet on the market side. It's a lot, a lot further behind in renewable energy side, we have a three, five and 10 year plan around that. And the legislation also needs to catch up, which will take a bit of time. Okay, thank you
very much. I think we have a question from Josh and when you've had so please just go ahead and then you can follow right away.
Yes. Just before the tour, a question from john. I just like to know who they really target client and your market and then how big to envision your market.
Yeah. So our our target market is molecule fuels and chemicals. And if you if you think about fertilizers, if you look at the chemical industry that can be which is currently from hydrocarbons that can be sourced from hydrogen and then fuel cell batteries. Which are also a very large industry for heavy fuel heavy vehicles. And then in the long run ammonia for shipping logistics, we did an assessment of the market and the market potential is more than a trillion dollars by 2050. And that will evolve over time.
Yeah. Yeah. Thanks. Thanks, Josh. Question you have mentioned AI as a service in your product. Can you a little little bit to explain now what is the value of your AI? At the end, actually, what was the exactly that the value that you are creating? And can you brief about the data sources that you're using for it?
Sure. Yeah. So the so the AI is a service is the the the the algorithms We will use that will connect to markets and to renewable energy. So to give you an example, we if you if you have a wind plant you we would be able to predict what you know what the capacity of that wind plant is how much kilowatt hours would be dispatched to the electrolyzer and at the same time be able to say, where would it go into the market. The other the other benefit we'd have is if the wind generator spins once and let's say generate one kilowatt hour, the system would choose whether to make hydrogen or to deploy to the grid because there's a demand. Let's say a person needs to charge their mobile phone or electric vehicle. So it looks at the best use case and the time they and what the up the optimal use of the energy is and that creates system efficiency as well. So so if you are a project developer or you're developing a hydrogen project, it would improve the cost of your hydrogen. And if you are a renewable energy project developer, it would find the best use case for your kilowatt hours that you are producing. So we think so so we think you would have a huge incentive to to work with us because we would already have a proven model. In terms of the data we collect globally from you know, guest network operators from companies that publicly publish their information. What we are working on is with grid operators and network operators in gas as well like in the UK, there is an initiative called gas goes green. So they are setting up a data system where they are going to supply data to the market. And we we are quite for that our, our uniqueness comes from our algorithms. So we are very happy to advance, you know, the democratic distribution of as much energy data in the world as possible.
Hi, I think I can I can ask a question now if that's okay. So thank you very much, because German is one of the questions that I had was where do you get your data from? You gave a very thorough answer to that from your presentation. Did I also understand that you want to build plants as well? Or did I misunderstand that?
Yes, we do. We currently we are currently looking at pilot projects in the UK, we have two pilot projects that we are looking at as part of the COVID stimulus and the in the economic zones. And one of the advisors that are here used to be the Director General of the pipeline industries guild in the UK and she is actively pursuing those opportunities. We also submitted to the Helsinki challenge, which is an energy challenge on heating. And we are we presented to the mayor and we hope to be shortlisted for that event as well for that challenge.
If I may a follow on question from that. If you are building plant that's obviously you know, capex and it's incredibly resource intensive. Are you funding that through grant money at the moment? And what will your long long term strategy be to make sure that that's monetized beyond the data? The data I think is the data aspect of your businesses is clear? It's the plants bit that I'm a little bit more. Yeah, I have more questions around Yeah,
sure. So So what will happen with the pilot projects is that we have relationships with established tech technology providers. So examples would be tasting crops, hydrate hydrogenics. And we would use their proven technology, but what we they are, they will be a use case for that plant. And if we prove the pilot plant with our technology, we do not necessarily have to go be part of the expansion of the plant, though it's an option. But we would be be quite happy to hand it over to maybe a much larger player, like a guest operator or international oil company. if, if, if that was the best case.
Thank you very much, much clearer. Thank you
very much, Ibrahim. I think that's it. Any other questions for you for now? representation? And I think we move forward with our next startup psychotic nudges from Spain. So I think you're gonna see us here.
Hi, yes. Ram
was good your packets from where we would see
Barcelona. Okay, how's this tuition there could be there better?
Well, it changes every day I would say. Yeah, the numbers are not quite good.
Well, I think we're doing our best. Stay healthy
and safe you your family, your friends and everyone I hope he stays ready. Okay, then what?
I share my screen
just for all the startups if you want to keep your video open positions to devise work. Yeah, the Tom is gonna start whenever you start speaking.
Okay, so what if I told you that there is a type of cancer with over 1000 new cases diagnosed per day worldwide has less than five months of life expectancy after diagnosis. It's the fourth cause of death by cancer in Europe. Would you know which one set or what are we talking about? Well, in this case it is the pancreatic cancer. According to recent studies in Europe, up to 25% of the population, which is around 100 and 50 million people have or will have at some point in their lives a pancreatic cystic lesion that might evolve into pancreatic cancer, but today, we can only do not diagnosed less than half of them. It is this current lack of precise screening method so that we don't have a screening method for telling where the cyst is and whether if it is going to be Malik, malignant or not. That is leading to unnecessary surgeries and lifetime follow up treatments for the patients impacting their quality life and the hospital's budget. Well, to overcome this problem, Ck technologies was born and we propose ck medical which is an assistant for radiologists that identifies classifies and predict the evolution pancreatic cystic lesions on any abdominal CT scan, we have two main objectives from one side increase the precision of the diagnosis fence to the artificial intelligence and optimize the follow up treatment. Thanks to our prediction model we've estimated we can save after 20 or 30% of these follow up image test. Our tool highlights the areas in which the syst our president proposes a probable diagnosis presents similar cases for evaluation so that it empowers their abilities, making a decision and provides a malignant potential by crossing the data clinical data of the patient with several past cases together get medically integrated in the hospital parks so that it can analyze the CT scan instantaneously even before it arrives to the radiologist for an analysis. Our business model is b2b targeting hospitals and health care centers in Europe. First, our product cost assess model software as a service with recurrent monetization based on a monthly fee. We target a wide market with standard product that it is scalable to other pathologies or organs and with our own technical development, which intellectual property is currently under protection, will in time in economical terms in terms of the market, artificial intelligence in medical imaging, is marketing, continued growth has done nothing but growing and it's about to reach to reach the 2000 million euros by 2023 and has a global annual growth rate of around 45%. Okay, and who are we who formed technologies. We're from one side Sara is a mechanical engineer as a Master of Science and MBA. She has over seven years of professional experience in Germany and Spain, and she has their CEO role. And Julia Rodriguez comas is a PhD in biomedicine postdoctoral research at the Institute of bioengineering in Catalonia, and he's our CSO, Julio, reverse is just anthropologies from the local hospital in Madrid, and is our clinical trial director and me myself. I am a robotics engineer, Master of Science also last year. PhD in computer vision. And I have the CTO role at our company. Maybe a pandemic had to come to engineer and medicine to work finally closer and benefit from each other. He's covered how much artificial intelligence can help in the analysis of abdominal CT scans, helping to prevent illnesses with problematic Indian, such as cancer, join the revolution. Join us at Sica.
And thank you very much representation. Maybe you're so maybe you've already enabled this video and join afterwards. We can have some questions or just a quick thing for our previous solutions. Some questions in the q&a or just for you. So also good to answer this question in the chat in the q&a. So please your back
of your Thank you. It was a nice presentation.
One question. First question is related to the your current status. Do you have revenue So far does that does the first question and the second will be related to the scalability Are you able to cover other problems or pancreatic issues is the only element where you are actually now focused on.
So we are quite a young startup. We are about to start our first clinical trials with real patients this month. These months are in September and so now we don't have revenue yet.
And the second question was related sorry, two.
Second question was related to the scalability are you focusing only on pancreatic issues or you are looking also for some other elements where diseases places where computer vision could assist you?
Yeah, so the technology would be the same right now we are focusing only in pancreatic cyst. Since we have seen during the ideation phase that it is quite a common problem for that alleges that it is not so easy. See to see. And we thought that this would be the first the best case as for us to start, but once this is mature we plan to move to other for example system, the liver or even other diseases that could be diagnoseable through a much in the pancreas itself. Thanks.
I have yet to very much that presentation. It was really, really good to hear and obviously you know if you if you do go through the clinical trials successfully, and then you'll be making a huge impact on people's lives. I have a question regarding the data and the images and the images specific to a type of person or agenda. You know it would you see different images if you had a man with pancreatic with a pancreatic cyst or a woman. And so so that would be one question. The second question relates the market I saw on your slide That you were targeting hospitals, but actually there were quite a few hospitals attached to insurers. So I was wondering what kind of conversations you'd you'd had with them in terms of testing their willingness to pay for something like this and integrate it into their, into the hospitals they manage. Thank you.
Yeah, so, um, in terms of data right now, so they met itself. It's doesn't matter if it is a male or a female, for example, but in our predictive model, the one that tells you whether if, for example, assist is going to evolve into cancer. It has into account several key factor from the clinical history of the patient, that we have also seen with our gastroenterology from the team. And, and this is the kind of distinguish or distinction that we do in terms of data between for the images and for the business model. Right now we are targeting hospitals first. We have had some conversations with other pharmaceuticals or insurance companies are so but as you were also commenting for it is very important that we have a successful clinical trial fair clinical trial first. So for them, yeah, they have some interest a hospital has have said that they would pay for something like this. But we have to assure that we have the accuracy that we say that we have in the clinical trials.
Thanks, john, I think you have a question.
Yes, please. Yes, sir. Very, very interesting presentation. And I like your technology. So I'm going to go into the technology itself, right and talk about data security. Okay. I'm very passionate about this area, because I'm involved in myself in a private knowledge. Okay, so with the with the data itself, so you know that you know, that there's a lot of data protection right now, especially when you go into medical. So how are you protecting the data? That's one question. And the data itself from each patient Of course, because a lot of these hospitals are not connected to the GPS. They had a lot of the GPS have their systems basically isolated under on computers. So how is the data being transported? second second question. So I have another question after that, but let's go to the first one.
Okay. So, yeah in terms of data, so, right now, we do have trained the algorithm with some public data is always anonymous, we do not have to would not need to have any kind of name or identification of the person. And then for the hospitals, the data does not leave the hospital for the inference. So to say in the training is done on cloud systems that are secure also for Okay, sure. And this we also work with another company that is also looking for these kind of things for us.
And on your algorithms itself, right. Are you using Are you using neural networks? Are you using convolution which kind of which way are you doing the How are you doing the modeling?
So right now, yeah. We are using deep neural networks for the semantic segmentation of the of the organs first and deceased. And so to say we predict its pixel of the image to its organ it may be. And then with the big picture, we say, okay, at least it was it assessed? Yes. Is it at the top of the pancreas at the bottom because this also impacts on the classification of it. So in this, we do it by deep neural networks,
okay. And under storage of the data itself, when you start when the data is stored, so cause, you know, you're you're taking a lot of CAT scans, MRI scans, and so forth. That's heavy data. Right. So is that all within the hospital network? Are you linked into that to the cloud? You mentioned? So so that's protected within the hospital? Correct?
Yes, yeah. So they also assure so it is also encrypted the data it is not something that it's easily also readable, but they also share this kind of protection.
And what's your outlook in one to two years?
Yesterday Diaz to get the European CE mark by mid of next year and start commercializing this product. And then the idea would be just to scale to a scale scale this to other pathologies, probably on the abdominal abdominal still. But this is in three to five years.
Okay. Thank you.
Yes, please be Joshua, go for his question. Then we can go back to Kenya. Yeah.
Yeah. So thanks for your presentation. And I think you have a really clear your problem, identification, and also your solutions and your technology pretty much clear on that you've got to really qualify the teams because I think it's really time for you to think about how to market your product. And then you said you're still young startup, but do you have any like no sales plans your marketing plan for the next year or otherwise, do you have any plan to get some I don't know, some fund from the investors or what they have any kind of monetization or sales plan.
Yeah. So we are planning to,
to go to a funding round by the end of this year, beginning of next year also with the results of the clinical trials since this will also pick up our product. And yeah, and this is currently the plan in terms of marketing plan and so on. We do have some things but as I said, we are still a bit young, and it is ongoing, I would say.
Tanya, she has another question for you. Good.
I do. I was just giving someone else a chance to answer again. So my question was, actually it was a phenomenon from you mentioned you're going to be applying for the CE mark later on. Next year, is there any other regulatory approval that you would need to seek before you are allowed to to commercialize?
So, in Europe, I would say the CE mark would be like the golden standard for us to look to go for it. As far as we know, so far and, and in the US there for sure. their, their version of this SP ft eight and so on. But in Europe, in Europe, this would be like for us our our mango. Okay.
Thank you very much. There was a lot of questions, but oh, we have one more from sorry,
I just have one more starting with quick one. Do you have an advisor? Do you have an advisory board? No. No. Okay. Thank you.
Great. So, I mean, I think you had a lot of questions, but great answers. And I think we have to move forward with our next. Thank you very much for being with us. Thanks to you. Bye bye. Our next step is from the US up your diagnosis. Are you ready to rock? Yep. Please unmute. Yeah. And on, put your slides up, and you have your boy.
Let me just set my four minute timer as well.
Okay, can I read? Can you hear me?
Awesome. So I guess I'll begin. So I'm sure that I'm the co founder and co author diagnoses and ocular we're trying to change the way we diagnose ADHD making diagnosis accessible, accurate and automated. So the biggest problem, especially if you look at it from a developed economic perspective, that equitable access to medical diagnosis is integral to achieve SDG goals in developing nations where poverty and a lack of medical infrastructure creates darkening qualities and the accessibility of diagnosis and treatment. Yeah, ADHD and similar neurobehavioral disorders are mean wildly under diagnosed, especially in the developing world as a result about societal stigmas, as well as efficient medical practices. The prevalence rate in Sub Saharan African Latin America is roughly equal to that in the United States and in Europe. Yet the actual diagnosis rate in Sub Saharan Africa and Latin America is much lower at issues of grievously neglected and extremely under diagnosed in the developing world. So this really boils down to three key problems. First diagnosis inefficient, it takes more than six months with wait time to one to two years. Second, and most importantly, extremely expensive. And basically, since it costs more than $2,000, it renders individuals who are in the lower end of the socio economic spectrum unable to be diagnosed with ADHD. And third, and most importantly, it's inaccurate. for someone to be diagnosed. They their teacher or their parent has to think they have ADHD. They go to doctor doctor is a piece of paper with a bunch of self reported symptoms. Based on the total number of checkboxes on a piece of paper. You either do or don't have ADHD and as a result of that diagnosis, extremely inaccurate 20% misdiagnosis rate and around 1 million children each year are misdiagnosed. The way we taught this at ocular is that we use people dynamic as a novel and objective biomarker for potential disorder. It's basically some research has been verified by global research institutions as as a proper metric of ADHD. And we basically rely on this objective metric as opposed to qualitative observations to diagnose ADHD. So it's really simple the way we work. First, a patient completes a visual spatial memory diagnostic application, which basically tracks their eye movement and the pupil dilation, then we capture this in real time using computer webcam. So it's basically accessible when it's on a mobile camera or a computer camera. Third, we basically use a machine learning front end or sorry, deep learning front end to to capture the people and process it and our machine learning algorithm. And then we output diagnosis as well as medical advice in case someone wants to crawl break. So this is a quick condensed diagnostic application. In reality, it's around 20 to 25 minutes but Basically, our password here, we can use this on any webcam. And then you can see over here, our deep learning on the front end is trying to extract the pupil of this test subject. And once it does, so it gives people it gives a visual spatial memory task. So this is just an abbreviated version. And then this would continue. And then once that is done, basically output a diagnosis. So no ages 2.2% risk, as well as biometric. So basically looking at the data and using that for a clinician. So our project Our product is really relevant for SDG, especially three in developing nation. The inaccessibility of diagnosis coupled with traditional stigma surrounding mental disorders, precludes proper diagnosis and treatment for you. So if you just think about it from an economic policy perspective, if we're trying to educate if we're trying to have education policy, there's no effect. If those children are being a not being diagnosed with ADHD, and they're suffering economically, socially. And academically. We'd go to The roots of this issue and diagnose ADHD as young as earliest ages. So education policy can be effective. And our team is diverse. We're basically all 16 or 17 years old, meaning we understand ADHD we have friends who have ADHD, we have family of ADHD. And we understand this problem. We're young, ambitious, hardworking, and the serial entrepreneur. And in terms of traction, so far, we've built our products were patented. And right now we're getting IRB approval. I'm doing clinical testing in local hospitals and pushing this into non local application we can visit as author diagnoses.com Thank you
very much. That's actually a super interesting solution. Same and any question from the judges and experts I'll do it for them. We're happy to send you a bit and your your solution as well. So please, don't
Like this screen is frozen here I can see cylinder presentation
Okay, yes, yes. Yeah.
It I loved it. I love the passion. Number one, okay, you got great passion behind you. And you're young, intuitive and it's fantastic. And the model itself that you have saw. So it is the business model in a way that this is for clinics or hospitals. Go to GPS, who's actually the your your first set of customer here, who's the first person you go to
the first business? Yeah,
so we're actually we're partnering with local hospitals and institutions that we've actually partnered with Mount Sinai, Icahn School of Medicine at Mount Sinai. So basically doing clinical testing through them. And in terms of actual revenue, it's basically going to be partnering with local hospitals institutional first and expanding that network and basically selling this to these hospitals. At the same time, we're also looking at partnering with insurance companies because if you think about it, for an insurance company, if it's a diagnosis of ADHD, the cost gets reduced so much as automated doctors not even necessary every day. As their costs, and they also have pathways into the medical into the medtech industry. So we're also looking to partner with them at the same time,
and already partner with any psychological institutes running on this and hdhd.
So we have a, we have a really renowned deep learning researcher at Mount Sinai, who is working with us, and advising us. And then also they have an ADHD department who's also helped us in terms of getting feedback and advisement.
And this application is our diagnosis itself, the data and MTS when you go through the process, this is all this is online, correct. This is online on an application, or is it a thin client that goes onto your phone? Oh, does it what is the
net format? So so right now what we're basically going to be doing is on our on the local application that we built that was done in the video that is basically a local computer, but basically we're going to be selling the IP the software as a software as a medical service to local institutions. Hospitals basically have that software running on their desktop or laptop or computer.
And you have this testing right now in some hospitals at the moment or the trial,
we're about, we're actually really close to starting clinical trials, the second round of clinical trials before we've already tested on 50 patients. And now we're gonna do like a full rounded clinical trial. Okay, super. Thank you.
Yeah, go ahead, even though maybe
chop, that that's a really lovely presentation. As john, I like the vision and energy. One question related to your business model, do you think that parents will be interested to pay and to measure the status of their kids and to get some some result property or because if that's possible, maybe your scalability of the business model could be even even greater.
Of course, to actually talk to patients, or sorry to parents, of children of ADHD. And basically the overwhelming consensus that we really gained from that the data that we understood is that parents are frustrated because for someone to be diagnosed, it's an extremely lengthy process. And the parent doesn't even know for sure. Is there cages there could actually eat you? Or do they not. So if they're receiving improper treatment, they're going to have side effects. If they're not receiving the treatment, they're not going to be able to succeed. So on the parents side, it's really just an overwhelming consensus that something is needed. And in terms of costs, the parents themselves wouldn't be bearing the cost is basically Well, it's basically automating the whole diagnostic process to changing the way we diagnose ADHD. So basically removing the doctor or just having them as as kind of a medical advisement. So basically lowering the cost for the hospitals as well as lowering the cost for the patients.
Okay, and who are your main competitors right now.
There's no one diagnosing ADHD out there. Except for doctors.
Hi, Sue. Thank you very much again, for the presentation. So I have two questions. One relates to a statement that you made and that one of the major problems is that there is social stigma attached in many geographies and in many cultures to having ADHD. So are you have you thought about partnering with other organizations that there's awareness around what this means and how this is not the fault of anyone, etc, etc. Because otherwise, I would imagine that the diagnosis, no one's going to want to have their children diagnosed as having ADHD because of the social stigma attached to that. So that's my first question. Then I'll move on to the second
court. So there's actually this nonprofit that I'm relatively close to, it's called vision For and from, I'm forgetting the exact name. But basically what they do is that they focus on kind of awareness in terms of ADHD as well as traditional neurobehavioral disorders. And they work for large people large, like a large population. So we haven't really made a lot of headway in terms of trying to reduce this. They might, because it's a huge issue. It's basically just the sidle issue that goes back a long way. But we are working with them in terms of seeing if there's any way just in terms of making marketing more positive towards these populations, or just making it so that more people can be diagnosed with ADHD. But I guess in addition to that, it's not just the Seidel stigma. It's also just people don't know about ADHD in the developing world, as well as the cost of actually diagnosing you have very few that you have very few people like neurologists in Sub Saharan Africa who have experience dealing with ADHD. So that's why we basically replaced the need to have those doctors and just use an automated diagnostic system.
Thank you. My second question is really a follow on from humans, which is the scalability of what you do sending into hospitals and selling into healthcare shoes since generally, I mean, the cycles, the cell cycles are really known. They're very complex, you have to go through procurement processes that may or may not result in something that isn't a purchase order for what you're offering, essentially. To what degree Have you considered even developing your technology, so it's something that people can do in their homes using their webcams? Is that something that your your technology could could address with a sufficiently you know, good resolution within a webcam?
Yes, so that last thing you mentioned is really like the biggest hurting block. So our objective is in terms of increasing the accidents, that's because of your application so that we can use it on any local application on any local webcam, but the problem is That just, it's really difficult in terms of giving a precise diagnosis on a mobile application without a presence of a doctor. And then also just a webcam, the quality of a camera on a flip phone or a mobile phone that has a camera may not be up to par in terms of being able to segment that pupil, as well as looking at it's my new site changes over time. But in terms of that, we're working to make our algorithm our deep learning algorithm using Hough transform methods, much more accurate in terms of being able to extract that people. So that's a vision. But right now for the short term. He basically talked with a couple of people in the medtech industry who run companies there, and they basically advise us in terms of FDA regulatory approval, and then also how to make headway into the med tech into the healthcare industry.
Thank you very much. Thank you.
So much sugar. I think we have Josh, final question. Yeah.
Yeah, well, actually, first Deliver thank you very much for your presentation and just like know what john said and I really love your passion and the seriousness and and I just quickly thinking about what I did when I was 1617. And so, I was really, really inspired by you know, your passion. So, thank you very much. So, some of the questions actually already covered by other judges. So, I just wanted to clarify my curiosity here. So, basically, you started with the issues about you know, inequality in treatment in the developing country, but also you mentioned about some, you know, problems in this treatment like an inefficiency like in a high price and also inaccuracy, but basically those challenges are still you know, also you know, applicable in the hospital in developed countries right so, you don't necessarily you don't necessarily really have no targeting to the hospital in developing country rather you firstly target to the any you know, hospital in any world and then you'd like to expand the your vision to the developing country. Am I correctly understanding your your vision
Yes. So the focus right now is in terms of we're based out of New York City for local hospitals over here. But the thing is that a big reason why the diagnosis is so expensive is that train neurologists, with PhDs, and a lot of research, and we're really good, cost a ton of money. And in addition to that, for someone to be diagnosed from that neurologist, it takes so many hours, it takes 70 months from your current visit. So that's why those bills keep on racking up. And that's expensive for hospitals even diagnose ADHD, and expensive for patients even come and get diagnosed with ADHD. So basically, by creating an automated process in which you're basically just having a royalty fee, it basically just makes the diagnosis itself much, much cheaper on the end of both the hospitals, as well as the consumers. So I actually just in my appendix I had a slide basically talks about how each year $3.5 billion are spent by households on ADHD diagnosis and 1.2 billion dollar paid by hospitals for the long hours of testing that we decrease the cost and increase the accuracy with AI.
Great, thank you very much.
Thank you very much representation. Also good answers for the questions and best of luck and thanks for being with us today. We'll move forward with our next startup coming from mentioned from Germany. So I think we have to vmg semester Yes.
Hi, how are you?
I can hear you as well. Okay, please just put the slides and make sure everything works.
And we have also Toby from same company TV energy. Hello.
Hi. Hello. Hello, everybody. Know
what's new for you start quickly with also tell a clear diagnosis that you have some questions in the q&a. be interesting to also check it out. And DFE please go ahead and do that. As you're searching, so speaking,
thank you. And Hi everybody. This is, as Amitabh is to be as Sebastian, we're calling in from Munich. And Sebastian, if you could go to the first slide, please. As you probably know, there are 1 billion people in the world without modern energy access. And you'll probably also know that that severely limits their everyday lives, their education, their health, and their aspirations. And yet, progress towards universal electrification is far too slow too far too few people get new connections every year. And a major problem is that the planners that investors in this segment in this market are shooting in the dark. There's just not enough data about the awkward villages to plan electrification. So they often don't know where the villages are, how many people live there and where the power grid reaches Sebastien. Next slide, please. With village data analytics, we want to make this we make these villages visible in a way that it's reliable. as fast and scalable, and we use AI to bridge data gaps, so we can tell companies or planners, what they need to know, we locate villages, we estimate their size, we tell them whether or not they're grid connected, where roads are where anchor loads. That's very important. And other factors, we can monitor sites, and we can predict the impact of interventions. So you can think of it as a Google Maps for the development world. Our technology is built on the application of machine learning to satellite imagery and on ground data. And my colleague, Sebastian will tell you now how it works.
Yes, thank you, sir. Yes. So our approach is this, we first identify the villages in an area of interest that the user defines. And then we extract relevant characteristics about the village that we identify in the previous step. And as a third step, we rank all of these villages that we identified according to user preferences in terms of how attractive are they to the user for off grid electrification. And this whole process is automated, and it enables a more reliable and more faster and scalable decision making for the user. And in order to achieve this, we use cutting edge algorithms from unsupervised learning and mobile networks that are trained on satellite imagery. And this data set that we get is up to date and reliable. And we obtain it through our data partnerships that we have, if you don't. And also these algorithms are confused because we see validated and improved through our work with the users on the ground.
Thank you. So I think we have a very good team naturally to deliver the data. And first of all, TFP, which is the company I founded originally, is has decades of on ground experience in building infrastructure for Rural Development markets. So we understand what holds the sector back and that's why we build them. On the picture there. You actually see my colleague Sam building a mini grid. I think that's somewhere in Kenya. Our technology partner is applied AI is Europe's leading nonprofit for the application of AI and brings in a very strong data science expertise. And then we're supported by the European Space Agency, which helps vieta with satellite data and with technical knowledge. Next slide, please. So we're currently working with a number of EDA users already that you can see on the left side design development institutions, governments and electrification companies. But we have built our technology for scale, so that we can expand it into different markets, and we can improve it with every project we do. So in addition to our current work and electrification, we already addressing a couple of new opportunities. The first is we're looking at new market segments. For example, health, this is driven by so we're currently supporting the COVID-19 response of one African country. And a second area is new geographies. We're so far we're working mostly in Africa. But now we're looking at first work in India and Myanmar. A third way to scale it is to add new information layers. We're currently working on improving the travel times algorithms. And the fourth is we're moving from offering a kind of single service to an ongoing monitoring service. And, you know, I stopped there. And I'm, thank you for your time. And I look forward to your questions which we can perhaps dive a little deeper into some of these aspects of our work. Thank you.
Thank you very much for respecting the timer. And I mean, you can have a chance to actually explain more in the future, the slides do the q&a. So yeah, let's go further questions and questions from explicitly. Yeah, I can see you around here. Please go ahead.
Yeah, thanks, guys. First question, just to be clear for me, you have cooperation with the company organization apply AI Is that saying that you don't have like internal AI capacity? That is developing your own algorithm? What is the relationship with the partners? That's the first question.
So, apply the AI has capabilities that we do not have in house and that doesn't cover the entire spectrum. So we have internal data scientists as well. But for the heavy lifting and deep dive work, we work with pi di and we have a long term partnership and market veida together as a team
Okay, and the second question, Who are your customers and how much you can earn from from them.
currently, the customers are working with development organ larger organizations, and of which they are A handful. And we're working with the investors into the electrification market the private sector investors and with electrification companies and and we can earn we can earn sufficiently from them we already are in Project revenues we are currently generating fuel our growth. So we can earn sufficiently from them to grow the market and grow in the market. But we're quite eager to explore other other markets as well. along, for example, health care or education, which has have slightly different sets of players, but it will probably in that in that developing context. And currently we're working almost all only with developmental players in the same market segment that we're looking at. We could potentially also work with agricultural companies, for example, and we could help companies that look at what understand value chains if they reach into these markets, they're different market opportunities. For us
thank you thanks
I to base and Sebastian, great presentation. Very interesting. A few questions for you it's so in relation to your product video, is that company that finished product completely? And
you are is a product ever completely finished? So we
have a very like for every participant is it finished in the product that it can be patented?
Yes. So it's it's it's a product that is currently in the market and is currently used and but as we are deploying the current version, we are building the next version
okay. So so you can be on better test and everything else and this is an American generated income.
We had a repeat
customers, okay. And you had you mentioned there, my colleague basically you know, he Ivana had a question you answered in relation to that somebody ei kebele is a The external. So if you weren't going to patent this in the future, right? How would you manage that? So what can you patent I should say, really? What can you patent?
it's an interesting question, but we're not so so at the moment I see our main value addition is in understanding the different data sources and data streams that are available in this area where we work and bringing them together in an in telogen way, and then creating a user interface that is really workable for our customers. And so it's a very application oriented approach is not so much a pattern of course, if there are papers that arrive from it, I mean, TT is a visa product. So as I mean as a TV product, so it would be with us, but that's not really the thrust of where we're going.
Okay, undefined one Last question. So your your focus in America. So this and to say the region's is this, this is not just only for non developing countries is also for developing countries correct.
So I yeah, it's a good question i, our our competency, I think our skill is in understanding rural environments in developing countries. That's where we come from. That's what we understand. And I think our technology, and what we do is, for the moment, and probably limited to that it doesn't it's not easily scalable to urban contexts in developing countries, and it's not easily scalable to develop countries, because the combination of on ground data and the patterns we recognize from on ground and satellite, they work very well these rural developing contexts. But if you look at a city or developing a developed country, then I would say the parameters of what happens in a place are many more. Okay, and you have collaboration with the European Space, the space associate Since Well, you said under satellite images, these images are data released see to pay for them, how was that I was that worked out. So, with the Sentinel imagery is freely available, okay. And and then in certain instances we work with very high resolution imagery as well, which is either which either we buy or it is made available to us by our partner. So a lot of the development institutions will have that and bring that into the project.
Okay, super. Thank you. Thank you. Yeah.
So, my question is, you guys said, you know, you have some customers already using your product data. Um, do you have any product what actually are they using your product, you know, as a pilot Aereo Do you have any pricing plan for them? For example, do you have like, subscription you know, They will see one time pretty smooth. So what's your actually pricing?
It's a brand? That's a very good question. The honest answer is that we were working on that at the moment. Currently, we sell it as a technology enabled service. And it works for certain markets very well. And it works for us. And the challenge for us, I would say, in the next six months or so, is to understand how to package it and price it for different types of markets. So that's and that's and that's a tricky thing, because but it kind of dovetails with our the increasing automation of the entire value chain that we're building.
Okay, just a second question. I know that there are quite a lot of false tech companies and startups are using you know, satellite imagery to the kind of similar project and how competitive do you seek in your domain mean? Do you have any like, competition related strategy to To make the other competitors, no failure to overcome the barriers Yeah,
yes. And so, a competition it depends on how you define your market right but tells you what competition is if we define the market widely as understanding rural villages in developing countries, then yes, there are a couple of companies they come probably mostly from the agricultural field that have very interesting technology. And what we do is we so by understanding electric electrification is the is a very, very good wedge into the village and ecosystem and the village economy, because it's often the first thing that comes in and so building a mini grid and an electrified village that unlocks all kinds of other things, whether it's modern agriculture, education, healthcare, all these other things. So understanding the village economy through through electrification gives us a competitive advantage. We have built a number of data partnerships with organizations that are working in the field in different countries in both in Asia and in in Africa. So that we can constantly validate and test this. So we've created a couple of case studies we use them to, to validate our predictions and improve our predictions. And, and, yeah, these are the main ways in which we can differentiate ourselves in the field of electrification. I think we currently stand out. I don't, I'm not aware of somebody who does the same thing. Okay,
well, interesting. Thank you very much.
Sebastian, thank you very much. I'm interesting that you're using the Sentinel data from from research to get that granularity of images. And so my question really is around your customers ROI. You mentioned a number of large corporates and one the slides they had actually there. For example, if they apply your Product your services if they purchase your services to really understand where these villages are, what the size of their markets might be, do you have any information on what the ROI can be for them?
And for building a uninflected, you mentioned LG, LG has a has different types of electrification products. It could be solar home systems, which are small products that are sold, and it could be mini grids.
And so so you're asking what the ROI of our customers in that market is? No, I'm asking for for a customer to purchase your services, what the ROI of their investment in your services. Do any evidence of that if you have any. Use Case studies in which you can say this company use our services and we were able to save the max because they didn't build or are we in a we opened up an X billion hundred million dollars. Yeah,
yeah, yeah. Okay, I understand. So, we've done that with we've done these case studies with a couple of companies. And and we have found that there are four areas in which where we can substantially improve their business development their business One is we can reduce the cost it takes to develop a new site by about 80%. Okay. The second is we can increase the time or accelerate the time or save the time it takes them to do that by at least 50%. And the third is that we can improve the bankability of their work and their investments by giving much more transparency on data and risks. And the fourth is that we can we allow them to scale which is the fundamentally the most important thing also in this market in order to achieve the global goals, and so that they can currently move from from developing tend to 20 sites at a time to two to 300 sites at a time or more. Okay, and that again, you know, if you come with these bigger portfolios, and they have no transaction costs that helps their investor ability and kind of the whole thing kind of moves ahead.
And do they typically engage with you then on a per project basis? Or do they? Is that a subscription model? Sorry, I probably
Yeah. So it's a currently it's a per project basis. And but that leads to what Josh was saying, we have different ideas of how we want to move that into a subscription, or a kind of an operational model, you could say,
and just a follow on question from that very quickly. Who is it that you engage with in these companies? Is it normally their CSR department? Or is it actually somebody at the coalface of their business development?
Yeah, so And currently, the maybe ology kind of through a little bit because he is on he actually runs a number of electrification businesses in Africa.
I've just seen so many
of these, we're always in. This is always that's the core business. These are companies whose core business is electrification. And, in addition, we are now engaging with impact funds and with companies that there are some interesting opportunities around carbon reduction and the different kind of angles to this that are coming in that might in more traditional businesses be CSR in more progressive businesses become part of a core business, just to thank you.
Thank you very much. Thanks. Thanks, Toby. Thanks. vestin repetition. A lot of questions. Actually, we're running over time here. So as time is super tight, you have to move forward with our next and last and definitely not least, startup. So thanks again, guys for the investment.
Thank you so much. Thank you. Bye Bye.
Thanks a lot.
And I think I'm happy to call on Gabriella for supima from Brazil. No,
no, no go. Maybe try to set your your slides and put your presentation on. And let all the attendees panelists and I mean startups experts know that we're probably gonna go a few minutes over time because we have only three minutes at the end of the session. So thank you other formats q&a, we might be like 10 minutes over time something which is which is okay. I hope it's, everyone's well, so Gabrielle, please go ahead. And once you start speaking, the timers will start.
Thank you. First and foremost, thank you for your time. Thank you for the opportunity to be here. I have a little bit of a problem with my video. So you're gonna have to excuse me for that, but I promise you are not that good looking. So I want to start the presentation with a simple question. rhetorical question, even. Have you ever been witness or better Part of a natural disaster in your life. And I believe for all of us that the answer to that is yes, we know that natural disasters are very dangerous and damaging to the society as a whole to our society as a whole. And with that, in mind,
supremo was born. Which,
which, sorry, I'm very, very sorry. Superman was born with the coalition of four main pillars. So prevention actions, they integration decision making support, and the surgeon equipment location. of our research, we discovered a more than free 2 billion block 140 $2 billion were lost, because disasters last year, and domestically more than 170 3 billion or lost in four years and over 100 60 deaths were occurring in Brazil. Other than that, there were structural and social damages and the OLC opinion on governments at the time was very, very decreasing. So we started to work on the following disasters thinking about that we start to work on defining disasters, which are, I mean disasters here in Brazil, which are inundation fluid lens light, floods, floods, storms, extreme temperatures, which affect mainly crops, you may see variation, droughts and mud flow. So how's the promo works? We have a system that is served to our to our customers via a subsearch a NOC subscription with monthly fees. And the system behind it works as three layered structure which is comprised of first entity To extract and normalize all data, second machine learning models to like, as I said, model the disaster itself, what are the main
components of the disaster or domain
risk factors of the disaster happening and the third deep learning system to try and predict how the disaster is going to evolve and behave over time.
a patient on our first pilot project. And this is this was made on the defense on the Civil Defense Department in the state of San Paolo. So, this is the first it's kind of ugly, I have to say, but this is the first pilot so supremely supremo AI, receives and collects data and money There's a specific region 24 seven, it can if it picks up anything any abnormal weather conditions, it already tries to predict what will happen in the next few hours. And it shows it prompt. It always shows it shows it probably to stakeholders which can make a decision send a squad or even communicate to all people who may be involved who may be in diarrhea, to evacuate or to follow instructions or instructions needed. To is a business part this is also part of our pilot. So we predicted in February of last of this year, a great peak of rain which caused me floods, I believe, over 600 points flood points on the brewery region, which is very close to our city tourism policy. And we predicted that with around 71% of accuracy. So our main strategy is to work with software and weather companies for data. We have our own hardware, and we also work very closely with them recently.
And I hope that I can just use a timer and again, you can also of course, Skidmore, your journey. So, I would ask now, our experts have any questions or anything of the note for this presentation. I can see john. There. All right. Yes.
Yes, very interesting presentation. You know, it's I think it's important that we have some predictive analysis, prescriptive analysis technologies there for certain disasters. But how are you plugged? Are you how expert question is How are you plugged into our what agencies are you plugged into to get the data so you can do the predictive analysis. And what kind of models are using.
I'm sorry, we were using mainly
ARIMA models or si Rhema models to account for seasonality or time passage. We also using regressive models have a myriad of reversing models on a on another layer. We have contracts we have contracts and partnerships with both academic institutions such as IPT and University of San Paolo and book
to gather data mainly together.
Specialized data. So SWOT analysis, when analysis from from the institute's and weather data from third party companies like Corona pymble and even know what? Okay?
And is the future goal of your application that it can be like for consumers that you can have an application on their phone? Or is it something that will be sold to telecommunication companies who is the end customer?
So we have two fronts on this. The first front major front is to be a B to G business. So selling to governments directly. We believe we have in fact many pilots with others, civil security departments around the metropolitan region, which is comprised of 39 cities of those 39. We are, we already are in 10 of those
We also have
a b2b segment, which of which would be mainly
mega partner logistic, logistics, businesses and
security assurance. Okay.
And how large is your company today? I know you're a startup, how many team members do you have?
We are currently on 17 members. We are very little startup small startup right now. But we are banking, thinking of expanding we need some other specialists on the case
as soon as possible.
Okay, thank you.
Thank you. Any other questions?
Okay, I see no other questions Gabrielle. So again, thank you very much. And actually with presentation, I mean, I have to say that a guy caught you on time, but I think it was clear. And maybe the only thing that was not clear with the team and then generally asked about it, and you answered, Yes. Okay, thank you very much. Thank you. And you. Actually, I want to thank everyone. I want to sync all the mentors and judges with maybe, although we were very tight on time and really few minutes over time, if I may ask all the other experts to enable the video and Mike and have a quick like 30 seconds, final remarks from each. And let me start with john here. So john, just a second.
I think just to give an intro To all of the different companies really kind of across the board. So I would really look into this because great to have a technology to have a technology. But look at the scale look at look at the scale of the model and who your competitors are and what you can patent and cannot patent from it. But also the funding your about your balance your books is extremely important, as essential. And if you're going for funding look at look at funded at a scale and scale basis only. So I would I would concentrate on your market with the market is make sure that you have a niche. Don't get too much. Sometimes we did get lost into the technology and everything else but it really I think, you know, think about the business because it's the business behind the tribes. That's my advice for all of the companies.
It's my right here in my view, Josh his remarks.
Oh, yeah, similar to Joan, you know, um, well actually really respected all The social entrepreneurs, especially who are really thinking about, you know, the, the, you know, solving problems in around the wall, and then also all the presentation. And so today basically they have all the really clear, no problem, you know, identifications and also, the idea of a solution for the solutions are really great. And then but also, as john said, I also suggest everyone probably better to have a really concrete plan for growth and scale up and you know, deployment offering strategies and etc. So because anyway, this is really business. So I think no, now I think everybody for everybody, it's really important time for them to think about how to really market their product and how to do it. We failed no monetizable in a business. So that's my last comment.
And you want to live video as well, but you certainly run for any marks.
Yeah, I think that john and Josh, actually, they said What I wanted to share here today, but I'm amazed really with the energy and the passion of the startup here today and full support from my side. I won't say anything more to that.
Okay, so very good customer to send you reminders with Tanya.
Thank you. So again, absolutely amazing presentations, very talented founders would like to see some women founders in that, that's just to it you next time, if you're going to have five presentations, maybe, you know, I find it difficult to believe there's not at least one talented woman out there that could have made it to the finals. So that's just a small thing. Not so small thing to say. And to the organizers. In terms of the what I would like to see from from the founders, if they're doing pictures in other fora, I think it should come across much more clearly, where you are at the moment. So what stage of development are you With your with your startup, what have you achieved, you know, give us some real headlines of things that you're really, really proud of. I'm not necessarily talking about money. In fact, I'm not talking about money, but we need to see some milestones, where are you? And really importantly, where do you want to be? And that that's my piece of advice if you're doing this in a different form. But thank you very much for inviting me. It's been a real pleasure.
No, thank you very much. It was very good point, though. The agenda point you mentioned, super important. I think some of the startups really have a, the one that we have today have their CEO, she's a female. I mean, I think technology's really, ever CTO today. So it's a very fair point. So thanks, everyone. And really, it was a pleasure. And I think the slides are closing in half. And now I'm going to show you the next sessions and problems we're gonna have for the next week. As you know, AI for Good Summit is now listed some bits all year long. So every week we're having a lot of elements happening and the pleasure to have you all Or the startups next week. Have a great week.
Thanks, everyone. Thank you.
Thank you. Bye bye. Thank you. Bye
bye is 11pm here so