In this Leadership Insights series video, Amanda Razani speaks with Francesco Iorio, co-founder and CEO of Augmenta, about generative AI solutions.
Amanda Razani: Hello, I’m Amanda Razani, with Digital CxO, and I’m excited to be here today with Francesco Iorio. He is the co-founder and the CEO of Augmenta. How are you doing today?
Francesco Iorio: I’m doing great. Thanks for having me.
Amanda Razani: Glad to have you here. So tell me a little bit about yourself and the company.
Francesco Iorio: Absolutely. My background is in computer science, more particularly in high performance computing. Essentially, throughout all my career, I’ve been working with very large computer systems solving problems that are very challenging, generally, for computers to do and even more challenging for humans, actually, to solve. And the company Augmenta actually is a reflection of some of that journey, if you will. We are using computers to help the construction industry change some very traditional avenues to design buildings that lead to a lot of waste a lot of waste both in construction materials and in energy use, right? So, by improving on the traditional design process, by adding a high dose of artificial intelligence – that really helps people make more efficient buildings to waste less resources, less energy, and overall actually increase the productivity of an industry that has been kind of lagging in general – the general productivity games that other industries have enjoyed actually, over the last few decades.
Amanda Razani: Wonderful. So let’s dive into that a little bit more, I know, you said that y’all used generative design machine learning, mathematical optimization and role based systems to create this new AI solution. So, how did you come up with that? Can you walk me through that and did you have any challenges along the way?
Francesco Iorio: Oh, absolutely. So, this methodology, if you will, that we that we created to solve this problem stems from the fact that the output, so what we produce are engineering design drawings, if you will. So they have to be, by law, compliant to rules and regulations that are determined by the specific governments rather than specific governmental agencies. So, unlike other generative systems that exist today that people can use and enjoy like you know, ChatGPT or Stable Diffusion that create beautiful pictures or some kind of enjoyable kind of text or chats, we have to essentially create something that goes under a lot more scrutiny if you’re will. A few pixels on a picture of a dragon in a fantasy setting could be wrong, and people won’t generally complain. If the designs that we create are wrong, there could be serious implications in terms of the regulatory compliance. So, we have to overcome a challenge in modern generative AI systems that is so called a hallucination challenge, right? So, these systems actually sometimes actually tend to very confidently report wrong answers to the statements or through the questions that are actually posed. So, what we did is we matched a traditional rule based design optimization system that is very kind of mathematically rigorous and correct, if you want, that kind of mathematically proved that the solution is correct. So, with the let’s say, speed and creativity of modern machine learning techniques, that again can be seen in these other systems that are available to customers to do this. So, we came to this conclusion, really, by matching are essentially deep background and expertise in both disciplines. So we had some some seminal research work actually in our careers inside not just a leadership team, but throughout our organization, both in the deep learning essentially section of computer science. So, but also again, in the large scale mathematical modeling and optimization that has been more traditionally the take, the focus of design optimization work.
Amanda Razani: Great. So what are some – can you share some use cases for some companies using this and then do you know, kind of, what the cost and time savings is using this?
Francesco Iorio: So the system is still in beta. So we are literally getting the product into the hands of customers effectively as we speak. But we see some preliminary results that are extremely encouraging. And so, again, some of the challenges I’ve mentioned earlier in the industry are, well the waste, again, up to 30%. People don’t realize but up to 30% of construction materials. that are used in the construction industry actually go to the landfill itself, you know, in the buildings themselves. So that’s a non-insignificant issue. So our goal and our preliminary studies actually indicate that we’re going to be able to save a lot of the wastes and to to reduce the waste by a considerable amount 20 to 30%. So which is really considerable, considering that this reduction actually comes from having more certainty about the building materials that are required for a project to exist and to be successful. So the level of detail at which our system designs, buildings, if you will, or portions of buildings – it’s such that, it gives a lot more confidence to people to purchase and to procure and to deliver on the construction site, all the the materials that are actually necessary, if you will. So while you know, we’re not going to claim that we’re going to reduce all landfills – that would be impossible, at least in the short term – but we definitely help in actually making that reduction. But more so there is a second kind of strategy, or if you will, the second benefit of using our systems in that the designs that are produced by artificial intelligence, usually, on average, they use fewer parts – up to let’s say, 10% less actual objects or parts to compose the system that performs the same function. So that has savings, of course. The cost, right? So of the systems that are installed, but of course, as a byproduct of having fewer parts, also the embodied carbon is reduced. Right? So just as a byproduct, using fewer resources, right? But critically, what our main prerogative actually is for the future is to design systems that are very efficient, so they use less energy. And so then what is required today, right? So, while we’re starting with electrical systems, where efficiency is only the same modest gains there – but in the future, essentially, we are going to create modules for our platform that also design mechanical systems, such as air handling, heating and cooling and ventilation systems, which are some of the systems that actually consumed the most amount of energy in buildings. So where the energy expenditure is the highest. And so having efficiency gains in those, cumulatively speaking, if all buildings actually work to be designed with our AI actually did, the savings actually will be really substantial in the future.
Amanda Razani: So definitely more green and better for the environment moving forward.
Francesco Iorio: Absolutely. That’s really kind of where we started, essentially, as one of the priorities of our efforts is really to actually create something that makes a considerable impact to the industry, both from an economics perspective, and from a sustainability perspective. And luckily, again, adopting this type of technology doesn’t have to be a compromise between the two. So it’s not always true that in order to make a system more sustainable, it needs to be more expensive. So we’re proving that.
Amanda Razani: And going back, you did mention code compliance – how do you make sure to be in compliance? And do you have any quality measures that you use?
Francesco Iorio: So compliance actually is very strict. So there is a very binary analysis -the system is either compliant or it is not. So we, let’s say, translated, the rules and regulations that exist into building code that are books that are specific to each country, and even to each municipality actually, sometimes, right? So we translated those rules and regulations into a systematic mathematical rules that we can enforce at design stage, and then check for compliance actually, at the at the output. So critically, we design our AI in such a way that we don’t have to check the compliance after the design has been created, but rather, the rules and regulations are part of the inputs to the AI, so that designs are created such that they are compliant by design, if that makes sense.
Amanda Razani: Wonderful. So I want to look at the industry as a whole and ask you how do you feel generative AI is changing the roles and responsibilities of engineers and other professionals in the industry?
Francesco Iorio: That’s an extremely good question. Actually. We think the injection of AI into these workflows, changes the relationship between essentially the stakeholders actually in construction projects, and will do so more into the future, mostly by essentially bringing the relationships actually closer together. So right now we see a transition and a shift actually in the industry from more – it’s a kind of segregated players right? So in for example, design bid build projects, right? So where the architecture and engineering team acts first and then after the bidding stages, essentially the contractors come in and actually design the building for construction and provide the necessary supply chain and then go forward actually into the built into the built environment. So, but progressively, we see the industry shifting into integrated project delivery, or design build types of projects, where all these actors essentially have a much closer relationship between each other, right? So we see artificial intelligence actually bringing them together closer, much closer, because of the speed at which the optionality can be evaluated or questions can be answered. Right? So something that normally takes weeks to answer for a question coming from a client in terms of actual space utilization, or sizing of the engineering material, right? So generally it takes an engineering team and or a contractor several days or weeks or months sometimes to come up with a design that can answer the specific questions. So generally speaking, this is not possible to do in iterations or essentially using the so called optioneering. Right? So essentially evaluating multiple options for each possible design choice, right? So by the introduction of AI, actually, that suddenly makes that possibility a reality. And therefore, the dialogue between all the stakeholders actually will be much more dynamic in the sense of decision making. Having the ability to very quickly examine, and, essentially, create options and viable options so that it really sheds additional light on the risks and the compromises that they could be making by making specific choices, right? So this is something that has never existed, really in the industry as an option or as an opportunity, right? So, but that suddenly its emergence actually will make again, these people actually, in the workflow actually play and work actually much more closely together.
Amanda Razani: So more options, better decision making, faster and more efficient.
Francesco Iorio: That’s right, and lower risk, right? So it lowers risk and lowers the costs. So increase the overall productivity really, and you know, while again, having a look at the impact on the environment, not be a secondary – an afterthought. But it’s really by design.
Amanda Razani: Absolutely. So thinking also about – some people have ethical concerns surrounding generative AI. Can you talk a little bit to those concerns? And how do we address those moving forward?
Francesco Iorio: That’s also an excellent question. So there are differences in the ethics concerning actually a generative AI, one of them is something that we hear in all different applications of AI, where AI can replace, sometimes, human skills and human creativity, right? So in our case, actually, the reality is slightly different than that. Design is a huge bottleneck today, right? In the industry, right? So and there really just aren’t enough trained and skilled industry professionals, who can take up the amount of work and the amount of challenges that the industry actually necessitates. So what we’re seeing in our system is future deployments actually, not in circumstances where organizations are trying to downsize, but are, in fact, actually, organizations are currently bottlenecked by the inability to scale because of the costs, but also, because of the lack of skilled trained actually personnel, which is really difficult to do have. The industry actually suffers greatly, actually, from lack of industry trained and skilled professionals. So we really, I mean, our name actually has that concept, like that very kind of built in sense that our aim is really to extend the abilities of humans rather than supplant them, if that makes sense. So AI, in this case, actually plays, we think, actually a very critical role in just essentially, really supplementing the creativity and the decision making power, if you will, that people already bring to the table, right? But also having the ability to create this vast landscape of options. It’s also learning tools that essentially people can use to really kind of better their own understanding of the compromises they are gonna make in each project that they actually undertake. So it’s definitely not a replacement, but rather actually, it’s a supplement.
Amanda Razani: Wonderful. Thanks for sharing that. Well, Francesco, do you have anything else to share with us today?
Francesco Iorio: No, I think that again, this is, we believe, one of the very few circumstances that happened to an industry like construction, right? So where there’s a chance for a transformation that is similar to you know, moving from the drafting table to a computer really, right. So it’s, we see the advent of artificial intelligence is really that milestone that will unlock essentially tremendous opportunities for the industry. So that really has suffered or lagged behind several others, like microelectronics or several others that have benefited from the injection of very sophisticated technology in their development.
Amanda Razani: Most definitely. Thank you so much for coming on the show and sharing your insights. I look forward to hearing from you again soon.
Francesco Iorio: Thank you so much for having me.
Amanda Razani: Thank you.