ResQ — Pitch & Mentor Q&A Techstars Startup Weekend Los Angeles: AI for Sustainability Soylent Innovation Lab, 555 Mateo St, LA Arts District — 21 April 2019 Full transcript, auto-transcribed from the original recording (MrVrAr-heads-up.mov). Timestamps are approximate. ====================================================================== [0:00] We are resting and to kick off our presentation, we have a short video followed by a presentation. [0:11] So let's not forget about the technical difficulties with distractions, the video, the lighting, but we have an audio overlay [0:40] as you can hear. Do you want to start it over? [1:10] So emergencies inevitably happen every year, house fires, adventure seekers lost on mountains, or trapped in caves. [1:17] And let's not forget about all those natural disasters that are frequently occurring around the world. [1:25] Like the current floods in South Africa and Mozambique, the mudslides in Santa Barbara, earthquakes in Northridge, [1:33] in these emergency situations, first responders need high-tech tools to help them locate survivors. [1:41] Introducing Rescue, a tool that helps firefighters and rescue workers find the holes in greatest need. [1:50] Even when they cannot see them, Rescue uses artificial intelligence to predict where victims are most probably trapped in rubble or in hiding from fire, [2:03] floods, or other natural disasters. The heat map displays this ever-evolving information in real time to a visor in the Rescue workers' line of vision, [2:13] allowing them to locate and save survivors more quickly and efficiently. [2:19] With Rescue, emergency workers have a new tool to save the lives of survivors and those who are doing the saving themselves. [2:29] Simply put, Rescue saves more lives. [2:33] So we're Rescue, and Rescue is an AI and machine learning technology that has been developed that is essentially middleware. [2:56] Right now, what we're looking at doing is simulating millions of scenarios in real time, identifying calls survivors, [3:03] displaying them in a simple heat map in visor, which are already technologies and hardware that are being used. [3:10] So simply licensing the data as a middleware into both providers, and then having machine learning aerial maps for command centers and operations and where to deploy their teams. [3:23] The problem is that first responder rescue processes are the same as they were in the 1950s, simply using radios and manual building schematics, literally going to the building walk clockwise. [3:34] Right now, the last five years, we've averaged 371 natural disasters per year, and due to climate change and other variables, this number has been growing at a rapid pace. [3:44] So has the average number of casualties, which over the last five years was a 66,000 fatalities per year number. [3:52] I also think it's an important number that we have over 500 million lives impacted every single year due to these disasters, and somewhere between 100 and 300 billion dollars in economic impact. [4:06] The inefficiencies of disasters on work and environmental disasters lead to greater casualties. [4:14] Our solution is to reduce time for first responders to rescue, simply creating better efficiencies through visualization software, [4:22] AI technologies partnered with the actual robotics of thermodynamic imaging, robots, drones, things like that, satellites. [4:32] Generate these things into an AR headset that the first responders are using, as well as intelligent tool for command centers. [4:41] Our goal, simply put, is to save more lives, to significantly reduce casualties, and we think that we have a measurable benchmark that's attainable to the tune of 10,000 lives a year. [4:54] That would be roughly a 20% efficiency increase through the implementation of AI and learning technologies. [5:03] Our low cost integration and middleware makes our go-to-market very easy. Our costs are pretty much covered. We don't need much to go to market. [5:11] Our competition are mostly partners because we simply layer in our data sets, and our revenue model is a licensing subscription model. [5:21] Also, worth mentioning government contracts. Last year alone, we spent $306 billion out of the US alone in natural disaster remediation. [5:32] The new definition of a billionaire, someone who positively impacts the lives of a billion people, we're really focusing on impact and letting our revenue fall out of impact, [5:41] which should happen in a healthy free market economy. [5:44] The market feedback is overall positive. 11 out of 11 market surveys said that they would open up their contact book and make introductions to NGOs and other needed partnerships. [5:55] The first potential client approach said they would not only anticipate this tech to be effective, reduce liability, but also reduce insurance costs. [6:02] So overall, there's a lot of budget out there for us to recapture out of the $300 billion that the US alone spent last year. [6:10] Our team is well qualified with extensive backgrounds in AI, machine learning, middleware, working, data science, nonprofits, and all the essential data and technology we need. [6:21] Thank you. [6:28] Question. [6:32] So I got a question. So did customers tell you what would be a significant amount of lives saved? I mean, how did you arrive at this number? [6:41] I also want to know what the insurance companies, what feedback did they give you about what their needs and gaps are in the market? [6:49] So working backwards, I actually own an insurance agency, do quite a lot of work in insurance. I think that the, if we can prove the first point, the second point will follow. [7:02] So I'd like to focus on the first point, which is measurable lives saved and damages remediated. [7:08] To speak to that, we simply work backwards off of that 66,000 lives died every single year right now. And if we could just make a 20% increase in efficiency, which the gap between the technologies that exist today and the AI technologies that we're implementing now, we think that that's a very realistic expectation. [7:34] That would put us at about 12,000 lives saved. So we'd like to make that our first full calendar year benchmark. [7:42] Can you talk a little bit about your go to market strategy you mentioned and NGOs, you mentioned first responders, like how do you imagine rolling this out and getting people trained on it so it's actually being used in real situations? [7:53] Right. Well, the majority of providers are actually concentrated into some real world giants. And so it actually makes it kind of easy in the sense that like for technologies, the headset that was really only one out there right now, Lex, one of our team members does most of their third part, does a significant amount of third party integration software for them already. [8:16] So this is a simple data layer to enter into the headset technology. In terms of data mapping and things like that, obviously, first responders like Red Cross and other major global NGOs are a great avenue. [8:31] I've reached out to one of my personal contacts who runs a 15,000 base international company that rescues a few hundred thousand people a year and has about half a million people and their direct impact with a few thousand employees. [8:48] She would be interested in actually using this in Mozambique right now because they're having problems in identifying survivors and some of the key members in the plains of South African Mozambique. [9:01] So our go to market would be to actually go to the large majority providers first and just simply offer subscription model that's very reasonable and prove the concept through the efficiency of those data sets actually pretty measurable results. [9:17] Speaking of the data sets, what specific data are you going to use to build these models to help rescue people faster? [9:25] Right. So firstly, it's real time. So that's part of the machine learning side of it. So every time that a person is identified, that's a new data point. Any time that the AR headset sees a human, that's a real data point, satellite footage of concentrations of people, data points. [9:42] Every time we walk into a new environment, we're going to have to go around and talk about it.