CUSTOMER STORY START-UP Deep Blocks: Using AI to Design Better Cities of the Future THE CHALLENGE When builders, architects, developers, and commercial brokers develop and design a building, they ve historically worked in a vacuum, unable to optimize their work against the whole project, causing delays and unnecessary cost. THE RESULT Deep Blocks developed AI-based software that combines, automates, and optimizes the work of multiple disciplines during the concept development and design phase of construction, reducing project time by 1000x and cost by 10x. COMPANY: Deep Blocks FOUNDED: 2016 INDUSTRY: Real Estate Development RESULTS, OUTCOMES, AND ACHIEVEMENTS Raised $1 million in seed funding 6 months after founding the company. 4 of 10 initial investors are real estate developers who tested the software before pilot. Process that took 3-6 weeks to reach optimal return on investment for a single property now cut to 15-20 minutes, a 1000x improvement. Only multi-discipline software to incorporate and automate the data and disciplines of architecture, market analysis, financial analysis, construction cost, and sustainability. Currently in 15 pilot programs with planning and zoning experts, architects, commercial brokers, general contractors, and real estate developers in Miami, FL.
She had an academic background in architecture and real estate development and had worked on big data navigation software. She had already been part of 10 other startup projects. Either the basic idea upon which they were based was flawed or they ran out of funding before reaching market. Each time, she learned lessons from their failures that she d applied to the next endeavor. To think that with AI we might potentially do that in a single system and be able to optimize it that s when my moonshot idea came together for me., Co-founder and CEO at Deep Blocks Now, she was looking to combine her academics with big data navigation but just wasn t sure how. At this point in her career, she was willing to try a new approach. The way she expressed her idea was unusual: I want to make cities work like forests and make buildings that are more like trees, each contributing to the larger, natural environment. So when a colleague returned from a Singularity University (SU) Global Solutions Program (GSP) that helps global innovators create moonshot innovations to positively transform the future of humanity, she decided to give the program a try. Her colleague had said the program had changed her life and her thinking about how to apply accelerating technologies to solving big problems. She is Olivia Ramos, the Co-founder and CEO of a very successful company called Deep Blocks, which she founded in December 2016 after attending SU s GSP in 2016, and later the Incubator and Accelerator programs, now part of SU Ventures. Deep Blocks is a software platform that uses machine learning to streamline and optimize real estate development analysis and return on investment. It is the only solution that dynamically integrates finance, big data, and architecture constraints and helps everyone involved in a building s concept development and planning work together to optimize their own contributions to the project....that s when I knew we could build a city like a forest, with each building one of the trees integrated into the whole, and at the same time answer the universal need for shelter.
In today s market, a complete development analysis is costly and typically takes 3-6 weeks to complete by 4-5 experts from architects and developers to municipal zoning specialists, commercial brokers, and construction engineers. With Deep Blocks, each expert can now produce the same results in 15-20 minutes, a 1000x and 10x improvement in efficiency and cost, respectively. How AI and Machine Learning Transform the Process Remembers Ramos, I was working with Bernard Leon, who was at the time my co-founder at Deep Blocks, at the GSP. I was laying out all of the components required to build a city one building at a time. He looked at it and just said, We ll have an AI do this. That simple statement was a life-changing moment. It blew my mind because without AI, the process requires so many different disciplines to pull everything together. To think that with AI we might potentially do that in a single system and be able to optimize it that s when my moonshot idea came together for me. That s when I knew we could build a city like a forest, with each building one of the trees integrated into the whole, and at the same time answer the universal need for shelter. We re not a blanket sort of AI that knows everything; we re using it very strategically. Before Deep Blocks, contributors in the ecosystem would use their own specific software with access only to the information associated with each expert s area of expertise. They worked in silos with little alignment, causing friction that slowed down the process and ramped up the cost. The first step for Deep Blocks was to change the software paradigm and allow contributors the autonomy to optimize their own pieces of the project. Today, Deep Blocks not only allows contributors to optimize their aspects of the planning and design, but also gives them access to pertinent data from across the project, so that when they optimize their own work they can see how it impacts everyone else s. Explained Ramos, If everyone is optimizing with all of the information, then the whole project is fully aligned and dynamically integrated, and the result is a better building. Once sufficient data is built and available to match to the specific pain points of the project, machine learning is added. We re not a blanket sort of AI that knows everything; we re using it very strategically, said Ramos. We built a product that could already do the work without AI, and then once we knew which specific calculations and datasets we needed, we used different tools within AI to implement them for a very, very powerful solution the world has never seen before. Sometimes we found it was easier to use other tools and just mathematics rather than machine learning. That s part of the learning process. AI is also used to optimize smart search capabilities and market predictions depending on very specific neighborhoods where the property is located. It is also used to adjust assumptions in what-if scenarios to optimize results. Sometimes we found it was easier to use other tools and just mathematics rather than machine learning, said Ramos. That s part of the learning process. As a young startup we wanted to use AI in everything, and we ve learned through the guidance of our Advisor Board member Neil Jacobstein, SU s AI department chair, how to use and implement machine learning the right way. Constructing a building involves multiple stages. Concept development and planning, the stage Deep Blocks helps with, is the first step. This is the phase in which developers or the project team look at a piece of property and try to understand what could be built there. Before Deep Blocks, the conversation was only about cost. But with Deep Blocks, the conversation is now about what should be built and how to optimize it. As a young startup we wanted to use AI in everything, and we ve learned through the guidance of our Advisor Board member Neil Jacobstein, SU s AI department chair, how to use and implement machine learning the right way.
The pilots are providing us with industry feedback from each vertical about what it means to be using a holistic software solution. Miami, FL: The First City to Optimize Building for the Future Today, Deep Blocks technology is deployed in 15 pilot projects with paying customers in Miami, FL. The pilots represent all of the disciplines and industries involved in the process. One pilot is for the public sector planning and zoning entity in Miami; several are real estate developers specializing in small to large properties; and others are commercial brokers, contractors, and architects. The pilots are providing us with industry feedback from each vertical about what it means to be using a holistic software solution, said Ramos. The process is simple. The specifics for each of the properties have been entered into the Deep Blocks platform. Two market experts then review the resulting report to ensure it is a good optimization for the property. The report, which is created nearly instantaneously, is then sent back to the customer. At this point, the customer has the equivalent of all of the disciplines on the team represented in the software s report on his or her computer. It is now available for the customer to continue to work what-if scenarios, which usually takes 15-20 minutes to reach the desired return on investment all based on realistic numbers and expectations of what the zoning allows. Deep Blocks is the first software in the market that is multi-disciplined and fully automated. Said Ramos, We re bringing in databases and populations from the industries of architecture, market analysis, financial analysis, construction costs, and sustainability. Eventually, our software system will analyze the entire building, including the passive and active sustainable systems, the construction breakdown and methods, and so forth.
The 10-Year Plan Once the design and planning stage is complete, architects generate floor plans and the shell design. Deep Blocks will incorporate floor plan optimization and space configuration in the software platform, as well as building code data for designing buildings in regional areas with coastal floodplains, hurricanes, or earthquakes. Eventually, our software system will analyze the entire building, including the passive and active sustainable systems, the construction breakdown and methods, and so forth, said Ramos. We re paying so much for the places we live in, and that s all due to how expensive it is to build in the first place. If we can reduce that cost, there s no reason why housing and the cost of living shouldn t be reduced as well. Farther into the future will be the addition of two other exponential technologies. In addition to machine learning, Deep Blocks will add mixed reality so that anyone analyzing a real estate project has an immersive experience to understand how the building will feel in the space. Blockchain will add the perspective of real estate investment to the development process, from understanding what will make a good investment and a smart contract to being able to provide a two-sided marketplace where the building is completely optimized and controlled by the software but players can come in and bid automatically in a secure environment. Here s how Ramos talks about the impact Deep Blocks can have: Think about it. Everyone lives somewhere, and everyone has the pain points of living in a place that doesn t optimize for natural light, for natural ventilation, for safety, for security, for economy. We re paying so much for the places we live in, and that s all due to how expensive it is to build in the first place. If we can reduce that cost, there s no reason why housing and the cost of living shouldn t be reduced as well.
The Lessons from SU That Made My 11th Startup A Success Olivia Ramos, CEO of Deep Blocks, had launched many startups in her career. All had failed to gain traction either because their basic idea was flawed or they had run out of money. She came to SU with yet another idea. This time the outcome would be different. Today I m applying everything I learned at SU. The most important thing I got out of SU was a continuous opening of my mind, in terms of the opportunities that technologies are offering to every industry today, and to some industries that have yet to become mature. It was one epiphany after another for the first month of the GSP. But it wasn t only a change in mindset to what s possible. It was also the way I learned to practically apply that mindset to what I was trying to accomplish. This was worlds away from what I thought a startup execution should be. The SU Ventures Incubator provided me with the first infrastructure for a startup, including all of the different pieces that you have to take into account all at once. This was completely unknown to me. I followed advice on every component diligently and I think this was the reason we were able to bring Deep Blocks to market. The GSP was about validating our moonshot idea and refining it for feasibility and for its ability to cause positive change against one of humanity s 12 global grand challenges (such as water, food, shelter, health, and environment). We also got to test our assumptions against trends that are coming. We learned which technologies would help elevate our startup idea to the exponential level. The Incubator focuses on how to get your idea into a 6-month plan to have a product that is relevant in the market at that time. At the Incubator I was joined by Udi Wolfowitz, who had also attended the same GSP as mine and who became the co-founder of Deep Blocks. He brought with him a background in machine learning and was the perfect fit as we began to scope and develop our concept into a minimum viable product (MVP). Once we had the MVP, which essentially means the product has been derisked and is ready to present to investors, we were invited to take part in the Accelerator and received our first seed investment of $100,000 from SU Ventures. When I think back about the impact of the Incubator on me, I believe I was almost blind in the world before it. The SU Ventures Accelerator helps you with the business aspects of the startup. An advisory board was curated for us and we met with many real estate developers, architects, and brokers to show them our demo prototype and get their feedback, which we then incorporated into our plan. For the first time in my various startup
journeys, I was able to define every single component of the plan. Beyond helping us test our assumptions, many of these industry players also helped us test the software and that is when some decided to also fund us. SU helped us with our presentation. We must have iterated it 30 times and presented to 75 different fund managers. But it was worth the hard work. Within 6 months after the Accelerator, we were able to raise $1 million from 10 funds. Some of the insights I gained that have proved invaluable include: Wait until there s a deep demand to invest in an employee or additional operations. You can t spend any money on something that would be nice to have. Because of this learning, we ve been able to build what we have and still have half of our funding from seed one in the bank, which has given us a lot of leverage to raise a second seed. This is one of the key insights that helps with achieving exponential growth. Perhaps one of the key mindset changes to result from my experience at SU was to understand the positive impact that AI can have. There is a common misperception that when you use these technologies you cause negative disruption. At Deep Blocks, we don t like the word disruption because it is often misunderstood to mean somehow replacing or cutting people out of an opportunity. Rather, what we ve built is a way for all of the existing industry players to be 1,000 times faster at what they do. We ve created a tool that allows them to be very efficient, to optimize buildings with far less cost. We re reducing the cost by 10x and the time by 1,000x. We re making these industry players exponentially better at what they already do. Design iteratively. Don t try to get it perfect the first time, as this is an impossible and paralyzing goal. Always be ready to throw away pieces, rework, and replace with new ideas. With this strategy, everything I learn along the way can be applied; with a seemingly perfect and therefore rigid system, there is no room for improvement. As a result, we have improved our software many times over and have come to a more flexible and scalable system than we could have ever planned for. Continue to interact with the market. We have a key meeting every week with a market player who gives us more information that, in turn, informs our software. In this way, you can continue to be as relevant as possible to actual users once you ve entered the market.
SU and Deep Blocks At-a-Glance SU PROGRAMS Global Solutions Program, 2016 SU Ventures (Incubator), 2016 SU Ventures (Accelerator), 2017 EXPONENTIAL TECHNOLOGIES AI/Machine Learning Data Science Mixed Reality (future) Blockchain (future) TRANSFORMATIVE PRACTICES Design Thinking Entrepreneurship Prototyping GLOBAL GRAND CHALLENGE Shelter About Singularity University Singularity University (SU) is a global learning and innovation community using exponential technologies to tackle the world s biggest challenges and build an abundant future for all. SU s collaborative platform empowers individuals and organizations across the globe to learn, connect, and innovate breakthrough solutions using accelerating technologies like artificial intelligence, robotics, and digital biology. A certified benefit corporation headquartered at NASA Research Park in Silicon Valley, SU was founded in 2008 by renowned innovators Ray Kurzweil and Peter H. Diamandis with program funding from leading organizations including Google, Deloitte, and UNICEF. To learn more, visit SU.org, join us on Facebook, follow us on Twitter @SingularityU, and download the SingularityU Hub mobile app. su.org @singularityu Rev 06-18 2018 Singularity University. All rights reserved. NASA Research Park Building 20 S. Akron Rd. Moffett Field, CA 94035-0001 USA +1-650-200-3434