From Concept to Carbon: Early Design Insights with AI and Autodesk Forma

ELLIS HERMAN: This is "From Concept to Carbon, Early Design Insights with AI and Autodesk Forma." We will, of course, start with the safe harbor statement. So a lot of what we'll talk about is forward-looking, is under active development. So please don't make any financial decisions based on the content of this presentation.

JACK RUSK: Hi, I'm Jack Rusk, co-founder at C.Scale. I'll cover in this presentation how embodied carbon data can be used to make decisions at the earliest phases of a project, and how AI helps fill in the gaps, supporting low-carbon decision-making at the very beginning of a project when data is scarce, but the potential for carbon reductions is high.

ELLIS HERMAN: And I'm Ellis Herman. I'm a product manager at Autodesk Forma. And I'll talk about how we in market Forma have tried to create a user experience that makes the process of evaluating embodied carbon easy and those AI models-- that Jack is going to talk about-- that power the analysis understandable.

JAY BURTWISLE: And my name is Jay Burtwistle, an associate with Stantec out of our Vancouver office. And I'll be building on what Jack and Ellis share with you and looking at a real world example. So it's a large scale redevelopment-- transit-oriented development project we're working on here in Metro Vancouver of about 20-plus towers that are going in on an infill commercial site. The architectural team has been using Forma extensively for the last year. And we are going to look at how we have then applied the embodied carbon tool to that project, so looking forward to it.

JACK RUSK: So to start at the beginning, let's look under the hood about how C.Scale uses machine learning for early phase whole life carbon estimation. But before we dig into the details of the models and how they work to drive the results Forma, let's describe the problem that we're trying to solve here a little bit first.

Typically, on a design project, 100% of the project data isn't available until the project is built. This is a problem for decision-making, since many of the most impactful decisions are made earlier in the process when relatively little data is available. This is especially important at the very beginning of a project when architects, engineers, and designers need to do important things like winning the project, and then go on to design the project again in a data-scarce environment. And often in the current workflows, it's not until too late that enough data is available to make low-carbon decisions.

The way our machine learning models work is they fill in data earlier in the design phases so that decisions can always be made in context, helping designers focus on what matters most. It's a noisy space. We get a lot of conflicting information from product manufacturers, consultants. But really, the goal of these models is to help focus on what design changes lead to the greatest carbon reductions, again, when very little data is available.

What ML allows us to do is fill this in, especially at the early phases when very little project is available. And in the project that Jay will describe, large scale zoning changes and project in the really most fundamental ways getting defined. So with that sort of problem statement articulated, let's dive into why we think AI and ML are an important contribution to this space.

First, we have to ask the question-- always-- of, is AI the most important thing here, actually? It's a buzzword we hear a lot right now, so it's important to ask a question, I think, about its appropriateness. Because not every problem is a nail, and AI is right now seen as a broad solution to many problems.

So to understand why we think AI and machine learning are appropriate for early phase low-carbon design, let's look at the two alternate extremes. First, we have the approach that's just taking a rough guess. And I use a picture of Kermit here because Kermit tries really hard to be green, but it's a struggle. It's not easy being green.

This is the approach where you use these magic numbers. Maybe the number is 350 kilograms of CO2 per meter squared. That really is essentially an arbitrary number based on gut feeling. And there's, most importantly, no parameterization. Well, if you start with 350, and you use low-carbon concrete, what does that number become? We don't know.

On the other side, there's really highly detailed generative design algorithms that design every column and beam in a building. These are based on first principles, often code-based analysis, using things like Eurocode. They're really highly detailed, but they tend to underestimate. Because of the realities of site constraints and program constraints, no building is ever this perfect.

But we view the power of AI and ML models in this space-- is having the right level of detail, having useful parameters that describe major design moves and material substitutions, and most importantly, having a basis in real quantities from real buildings, allowing early phase predictions to be informed by real outcomes on real projects rather than hypothetical outcomes from idealized projects.

So if the process of using these models is going from a very low resolution image of the building at the early phase to something high resolution enough-- again, still at the early phase to be able to make a decision-- how do we do it? And the answer to this is our machine learning models which, thankfully, aren't unique because people have been using them for decades at this point to look at pictures of cats on the internet. And these models that we use do actually very similar thing to what the models that you might encounter in other image generation models do, where we take something at a relatively low resolution and build resolution within the image so you have a full picture of the building at the beginning of the process. So now, I'll walk through the flow of those models and how that data is generated from at the early phases.

The first step is parsing the input data. So that's whatever the geometry drawn in Forma is turned into a set of predictors that describe that building in a way that can talk to the background data sets and models that are being queried. Some of this is very straightforward, like number of floors. Some of it is synthetic data that we create. Like the volume of a building is an important parameter. And this structured set of predictors is matched to those used to train the background data models made from data on real buildings.

So once we've done this translation process from what's drawn in Forma to the format that can be used to query these models, the first quantity we estimate is total SMQi. SMQi is Structural Material Quantity Intensity-- the amount of structural stuff per meter squared, or square foot, of a building. And this is a predictive model based on about 4,500 bills of materials from real projects.

This model is able to estimate about 3/4 of the variation in the underlying data set. This means too that the model misses things. Very small buildings, especially when designed by architects, tend to be idiosyncratic. Very tall buildings, which we can talk more about in a little bit, and buildings that are structurally expressive-- might have large cantilevers-- aren't going to fall into fat middle of buildings that we're able to describe really well with these models.

So once we've estimated the total Structural Material Quantity Intensity in the building-- the amount of structural stuff per area-- we then partition that out, adding resolution to the picture of the building into major material categories. How much of it's concrete, how much of it's steel, how much of it's wood, reinforcing, blockwork? And this model performs exceptionally well, capturing about 95% of the variation in the data set.

But there's relatively less data to support it. So we have about 500 bills of materials that allow us to do this level of partitioning. After this, we partition it a step further, adding additional resolution to the model. And this is where we go from concrete as a category to concrete of different strengths, or steel as a category to cold-formed steel and hot-rolled steel.

And this gives us a pretty good preliminary bill of materials from the building that we are able to infer just from the geometry drawn in Forma. Able to do this entire process of adding resolution at the very beginning of the design when decisions about the building are still being made. But this material quantity estimation is actually only half of what goes into the model. The other half is the background carbon intensity data.

So rather than looking at just one product at a time-- because at early phases, these supply chains often aren't determined, and material sourcing is still an open question-- we look at the range of data available for each building element using commonly used EPD aggregators, like Eco Platform, EC3, and generic data, like that available through Quartz and Okobaudat, including other global data sources depending on the region that the building is modeled in. From this, we establish a distribution of all the materials available that could meet a certain specification, and use that to sample carbon intensities at the 20th, 50th, and 80th percentile, giving a idea of the range available and the amount of carbon reduction that's available by improving specification. Currently, in Forma, the default value is the best practice-- is 50th percentile value. So with that understanding of the background, I'll pass it off to Ellis to talk through In-market Forma.

ELLIS HERMAN: Thanks, Jack. So again, I'm going to talk about how we at Forma have tried to create a user experience that makes the process of evaluating embodied carbon more accessible and those AI models-- that Jack talked about-- that power the analysis more understandable. So very quickly, what is In-market Forma? Forma is a fully cloud-based early stage planning and design product, where every project is geolocated, which means that it's super easy to pull in a lot of that contextual data-- surrounding, existing buildings, terrain-- with just a couple of clicks.

Forma has easy-to-use drawing and editing tools in addition to interoperability with Revit, Rhino, which let you quickly experiment with a lot of different building forms and site layouts. And most relevant for this presentation, Forma's biggest goal is to enable outcome-based design by providing easy access to those outcomes while requiring little to no technical training. So because all of these sites are geolocated, all you have to do to run a sun analysis is enter the date that you're interested in. Wind and microclimate have all of that local weather data automatically sourced.

So again, very quickly, how to access Forma-- again, it's a fully cloud-based tool, so there's no download required. Many of you probably already have access to it because it's included in the Autodesk AEC collection. There are a couple of URLs on the screen here. You can also just Google Autodesk Forma, and it'll be there. There's also a 30-day free trial available if you don't already have access. So over the last year or so, we've been working with Jack and C.Scale on building an embodied carbon analysis and Forma powered by that AI model that Jack described. And over roughly the same timeline, we've worked with Jay and Stantec to understand how to create that user experience that makes the analysis accessible.

And over that time, we've talked to many different Stantec practitioners-- architects, engineers, sustainability specialists. We've asked questions, we've tested these designs, we've tested prototypes. And now, we're actually showing the real In-market product being used on real sites.

So through all of that work with Stantec, we got a much better understanding of the current problems and bringing embodied carbon into the early stages of design, into making it one of those outcomes around which you design. So we learned that most of the time, embodied carbon analysis is accessible only to carbon experts who are not usually the ones driving those fundamental massing and primary material decisions. And that because of that, these outcomes are often not considered until too late in the design process, which means that we miss out on the chance to affect those most impactful early stage choices.

Now, all of that work led to our presentation at AU last year. And at the same time as that user testing was going on, we were starting to work with Jack and C.Scale on understanding how to integrate their model into Forma. So from the beginning, it was pretty clear that our natively drawn Forma buildings fell somewhere on the left side of this graph that Jack showed earlier, where users have defined a basic building form but not much else.

So by asking users for a few primary material inputs, we could add enough detail to start making reasonable predictions for the resulting bill of materials and thus embodied carbon. So here's the demo. You'll see we ask you to input a building program, a couple of cladding parameters, and a primary structural system.

And our goal here was to balance accessibility, making it easy to get started and run this analysis for as many ideas as you have-- with getting enough specs about your building to make useful, specific, actionable predictions. And in this first implementation, we leaned towards accessibility. So the analysis takes just a few clicks to get started. It runs in just a few seconds.

And this is one more reminder that Forma is a fully cloud-based product. So we're able to continuously push improvements to our features and really take into account your feedback as we go. And what that means is that, in the coming months, while we'll work to maintain that accessibility, we'll also add more transparency and customizability of the data behind the results that you see here.

So what should you use this analysis for? First of all, for informing those early stage decisions. Forma right now is primarily aimed at architects doing early stage design. We want to enable everyone, regardless of carbon expertise, to understand the relative impact of different building forms and primary materials.

Secondly, setting goals. This is a whole-building embodied carbon prediction so you can understand from day one how your concepts compare to baselines, to regulations, to firm commitments. And thirdly, to starting the conversation. We were reminded over and over that conversations with clients and other stakeholders are a driving force of early stage design, and that persuading those stakeholders to opt for lower carbon options requires real numbers and real evidence.

So with Forma's embodied carbon analysis, we hope that we can both enable architects to perform this kind of analysis themselves to lower that barrier to entry and bring them closer to the sustainability consultants who previously often entered the conversation too late. So like I said, we've been working with Stantec practitioners for the last year or so. So it's very exciting for me to pass the mic to Jay to talk through an actual use case, where Forma and Forma's embodied carbon analysis has been used on a real project.

JAY BURTWISLE: Perfect. Thanks so much, Ellis, and thanks to you and Jack for all that you guys have done to bring Forma to market and allow us to test it on this project. So for those that maybe don't know Stantec, we're a global firm founded in 1954 in Alberta, Canada. And we've since spread all over the world across 450 locations, and we're up to 31,000 employees now. So just a few of us around here. You might meet some of my colleagues at the conference.

And we've been working with Forma for quite some time and also deeply concerned with carbon for quite some time. So we've really been focused. I'm part of our Western Canada carbon impact team, and we have carbon impact teams across the firm. And it's really a recognition on the part of the firm to address climate change as much as we can on our projects.

It's an acknowledgment that we can only change what we measure. So being able to really quantify carbon on our projects is very important to us. And we are signatories to Architecture 2030 and really striving to reduce our carbon on every project that we're working on.

And it's also in response to clients-- where we're seeing increasingly requests from clients that have committed themselves to reduce their own carbon footprint through ESG commitments or regulatory policy commitments at different levels of government-- that we're working with. And similarly, those levels of government are also introducing regulatory requirements that are starting to drive the conversation around carbon in the built environment to ever increasing levels of rigor.

And personally, why I care-- I'm a self-professed regenerative junkie. Very much care about the built environment coming from an architectural education background, and really wanting to appreciate sustainability through a holistic lens. And seeing life cycle assessment, I think, as one of the ways that we're able to quantify the built environment's impact on the whole life cycle of a building from extraction of materials straight through to end of life.

And so through that, I've been consulting as a sustainability consultant for the past 11-plus years in a variety of settings by a variety of different sectors and scales, and have been doing a lot of LCA analysis. But I'm a first time Forma user. This project was my first time getting my hands wet-- my feet wet with Forma. And it was really, really great to be able to dive in on a project.

Some of my past experience from doing that LCA, I think, really identified a lot of pain points in the process, similar to the journey we went on when energy modeling was first introduced into the marketplace by LEED and other requirements. It was very much a linear process. We saw the architect would come up with a design-- either pre-design, concept, schematic-- would share that design with the different disciplines who would kind of resolve or further evolve their different disciplines' aspects of the design-- mechanical, electrical, structural, et cetera.

And then that information would flow to the green team, the kermits of the world. And we would really then have to take all of those inputs and run an analysis. And in the case of energy modeling, sometimes they would go away and take two months-- or two weeks, a month, something like that to produce results. In the case of LCAs, it's been a very similar process.

I think we would get that bill of materials-- the quantities of all of the materials in a product passed to us from the team, essentially. Or in some cases, we'd have to do take offs from the drawings ourselves. We'd have to make several assumptions based on past project experience or marketplace standards of practice of construction and then input all of that into some LCA software.

At Stantec, we've been using OneClick LCA for the past several years, but also have experience with several others on the marketplace-- Athena, Tally, et cetera. And then we would build out our baseline based off of that and then run these sensitivity analysis options to say, if I reduce my concrete by this much, if I adjusted this cladding material for this other cladding material, et cetera. And all of that would take place in this green box, or black box, where really, we're the ones that are running these analyzes and interpreting these results. And it might take us a week or two to go through that whole process.

And then we're able to share results with the team, have a discussion. And maybe we would iterate that several times. So it is an iterative process. And it does involve a conversation with a team. But from a scope and timing perspective, we really have to get that information, take it away, do our work, and then share the results.

So I think the exciting developments with Forma incorporating now this embodied carbon analysis is that it allows for much, much more dynamic interactions among the architects and the rest of the project design team members. So really, we can sit together and look at the same model. Because it's cloud-based, we can do it remotely even, and really look at, if you make these design decisions, this is the impact on the embodied carbon. We can run the analysis. We can interpret the results live and then run some iterations.

So we'll go over some of what we've done on the sample project here. But you can see at each step of the design process that it would be this iterative process, where we would all sit down together. And then schematic design would evolve to design development, would evolve to construction documents, and so forth. So it's quite exciting, I think, to see.

And it was quite exciting to see on this confidential project. Unfortunately, we can't name names, but it's quite a large scale. You can see here from the context in Metro Vancouver. This is following a pattern across metro Vancouver of these transit-oriented developments happening. I want to say there's at least six going up right now of these multi-tower projects around our SkyTrain or sub-- other cities have subways-- we have a SkyTrain.

We like to fly. I don't know. So the developer approached us and really wanted to explore what was possible on this site and to begin the process of looking at master planning across the site. And we developed a package and looked at some renders to show them what was possible, introducing a new park and several towers. So it's about 25, 26 towers proposed in the current iteration.

Most of those are residential with a commercial or retail component on the ground floor, kind of typical to many places, I think, around the world right now, as well as there's a hotel component and things. You can see here the SkyTrain in the foreground, and new public amenity space, and all of these kinds of things. And the team went through and did quite a bit of analysis looking with Forma, looking at what was possible from different massing, different height options, looking at the impacts through sections-- so being able to pull all of these early concept drawings straight from Forma-- looking at Shadow studies-- some of what Ellis talked about-- being able to do solar, and wind, and these kinds of things.

And then Jack and Ellis came knocking on our doors and said, hey, we've been working on this new tool for embodied carbon analysis. And we said, that's great because we've seen-- here, we have an excerpt from the City of Vancouver's climate emergency action plan-- where we've seen this trend in our region, at least, as well as many others of municipalities and other levels of government really pushing embodied carbon to the forefront. As energy codes and things have driven the conversation around energy efficiency and reducing operational carbon, we've seen a real, really strong shift in the marketplace to be able to understand and reduce the embodied carbon of our building materials as well.

So here, we see Vancouver, for example, has committed to 40% less embodied emissions from new buildings and construction projects compared to a 2018 baseline by 2030. So their goal within the next five years now, basically, is to dramatically reduce the amount of embodied carbon in our new building projects from that baseline.

And to do that, they've introduced embodied carbon guidelines-- how to create models and all of these kinds of things. And we're seeing California, Toronto, lots of other jurisdictions adopt similar policies and similar guidelines to actually guide teams how to document. And it was exciting. I think Jack shared just a week or so before this presentation that the City of Vancouver has recognized Forma as one of those early adoption tools that you can use to test at the early stages of design to confirm whether your project is complying with the City's policies.

Yeah, so another just great reason to use the tool, if you're working on projects up here at least. But what we've seen also is that, actually, the city of Vancouver, at least in our region, really leads the way. And a lot of other municipalities end up adopting some of their guidelines to roll out practice and policy in their own jurisdictions.

So it was this win-win scenario of, we've got a project that we're working on. It's not in the City of Vancouver, but it's in the region. And we've got a tool that we can test to see how we can reduce carbon on the project. So we dove in. And as Ellis mentioned and demonstrated how to select those parameters on the right hand side through the embodied carbon tool, we went through what we would normally go through on a embodied carbon analysis and a whole building LCA.

On this case, we're looking at a whole neighborhood LCA. But we built out our baseline-- looked at a hybrid concrete structure for high rise as our assumption. We looked at a GFRC cladding as a typical practice in this region, and 50% window-to-wall ratio. It's a little bit higher than we might see based on some of the energy codes, but it's the default in the software. As well as for residential projects, we tend to see a higher window-to-wall ratio, so we thought to keep that for our baseline at least.

And then as I mentioned, it's a mixed use residential and commercial. So we set all of-- the majority of the buildings to a residential use. We were able to select the way we had built up the model, select the podium separately, and set those to a commercial use as well as the one-- there's one commercial tower at the center, and a hotel which we set to a residential use. So set all of those up.

Diving in as a first time Forma user, I think I was given access to the model. And making all of those adjustments took me maybe an hour and a half-- something like that-- to get familiar with the tools and go through and select the buildings. You're able to select multiple buildings and change the features.

So it was pretty quick. There was a little bit of troubleshooting, which I'll talk about in the lessons learned in terms of some of the geometry. But generally, it was a pretty quick process. And then as you can see on the right-hand side here, it's 5 to 10 seconds to run the analysis. So we're able to click on that button, and we get our results.

So this highlights in the color gradient the relative embodied carbon impacts of all of those buildings. And we can see-- I've blown it up a little bit here on the right hand side-- the results. So the total embodied carbon for all of the buildings-- 723,000 tons of CO2. So that's a not insignificant amount. Given the scale of the project, not too surprising.

And then the average carbon-- so this is what sometimes we refer to in the industry as well as carbon intensity, if you're not familiar, which is a per square meter evaluation. So it gives us an ability to compare buildings of similar types to say, per square meter, this building has a higher carbon intensity than this other one. So again, across the whole site, through all of the buildings, we see, we have an average carbon intensity of 409 kilograms of CO2.

So this goes back to what Jack was saying. Sometimes we just throw out these numbers, and you're like, well, what does that mean? At this early stage, there's obviously some level of information missing. For example, in our model, we didn't have a foundation. So the number is a little bit lower than-- this is where having a sustainability consultant step into the conversation is helpful just to highlight some of-- from experience, to say, well, this number is a little bit lower than I would have expected on a large scale high rise residential and commercial building.

One of the main reasons is probably because we don't have any foundations or subgrade structure included in the model. And so that was a big learning, I think, for us was like, obviously, the foundations on these large towers would be a pretty significant portion of the embodied carbon. And again, I'll touch on that in some of the lessons learned.

As a baseline, it gives us now a number that we can compare to. And that's really where the power of early modeling for embodied carbon is so useful to teams to be able to say, even if I know-- and I know some of these assumptions-- I'm still able to use this model to compare. If I substitute materials or look at different design options, then I can say, relative to my baseline, this version is performing better.

At least in our marketplace, one of the immediate knee-jerk-- it's almost becoming a knee-jerk reaction now for reducing embodied carbon-- is to look at mass timber as a structural option. Currently, in the province of British Columbia, we're allowed to build up to 18 stories out of mass timber under BC building code. None of our buildings in this development are 18 stories, so we to-- for learning purposes, basically, we said, let's look at the shortest few towers assuming that building code in the province might catch up.

Elsewhere in the world, they are building taller out of mass timber. So we expect that in the future, building code would catch up. So let's say-- let's assume that for the shortest building-- so we have these three I've highlighted here in the lower-left-- are the office tower actually, and one residential tower, and the hotel. And they're 35, 40, and 28 stories respectively.

So we said, let's assume that in the future, the building codes catch up, and we're able to do these out of mass timber, so let's change the structural systems and Forma and see what our relative impacts would be. You can't quite see it. I didn't highlight it here, but the average carbon for just these three buildings is 313, so quite a bit below our total average across the whole site.

But it's exciting. Now, we can actually see across the whole site how does just changing those three buildings impact our total embodied carbon. So we see the results here at the bottom right. The total tonnage has dropped to 681,000 tons approximately. And the average carbon, which is the carbon intensity, has dropped 394 kilograms. So about a 15 kilogram per square meter drop across the whole site just from shifting those three towers.

And if we look at the total carbon saved, it's just over 42,000 tons of CO2 saved. So it's quite a significant-- I mean, we're looking at a very different scale than maybe a single building if you were doing this analysis. But you're able to, I think, really appreciate the measures-- or the significant impacts that some of these design changes can have at an early stage. So it was really great to be able to dive in, and use the tool in this way, and start to have these conversations with the design team. And the client is also very excited to hear some of these results. So we'll be sharing some results with them soon as well.

And then we wanted to look at another use case as well for Forma. So we looked at as well several other sense what we would call "sensitivity analysis--" that I talked about of like swapping out cladding materials or changing the window-to-wall ratio. I didn't want to go through each of those for the presentation just for time's sake. But I thought also just to show this other use case of actually looking at the built form. And I think this is another area where Forma can be really exciting.

And so we just zoomed in on this one tower. You kind of see the L shape. And most of the buildings on the site are fairly rectilinear, so we zoomed in on this one because it was a unique geometry. And we said, let's just see if we made some changes to this geometry, A, how easy is it, and B, what are some of the embodied carbon impacts?

So I just have a little video here to show what that looked like. And this is live. The video is two minutes. And within two minutes, I was able to run two different options. So the first one, I said, if we change this geometry to more align with the street-- the street here-- that's why that the building is kind of shaped that way.

But I thought, let's bring out, that outer edge to more line up with the street instead of with the podium or with the other buildings. So less rectilinear. It's a bit of an Arrowhead shape, maybe you would call it. And so made those changes-- was able to run the analysis very quickly. And boom, results.

So we could see a slight reduction in overall carbon. Very minor, but the point here is you're able to see really how quickly we're able to run these analysis. I think. And then similarly, I thought, what would be another use case that architects might want to do? Let's reduce the height of the tower and increase the floor plate. So trying to keep the overall, floor area roughly the same-- number of units, et cetera, that the developer would care about. But maybe reducing the height so that we have more allow more sunlight and greater floor plates.

Again, was able to make the edits to the geometry within a few seconds and run this analysis on this building. So you can see that I could sit down with the architectural team very quickly and say, if you're thinking about some of these options, let's run some of this analysis. And within two, three minutes, provide some level of feedback on the embodied carbon impacts of some of the design decisions that they could contemplate.

So it's quite exciting. I think we learned a lot of lessons, being that it was my first time. And Jack and Ellis were amazing and kind of being responsive. And as Ellis touched on, they're constantly pushing updates. So even I think out of this process, we identified several that they've already implemented in terms of the user interface and a couple of other things.

But just to highlight a couple of the main ones that we learned out of this, modeling is important. So we ran into some clashes in the software of overlaps. So we had geometries that were conflicting. And we had to go through that. Part of that first hour and a half was just kind of fixing some of those geometries.

And the foundations, as I touched on, we didn't include any, and that would have a pretty significant impact. So I think model, if possible. So include some subgrade structures even if it's just an assumption of, we're going to have 10 stories of underground parking-- in a tall residential tower, we're going to have three, we're going to have none, but we're still going to have some foundation. So somehow incorporating some level of subgrade structure, or make some assumptions if not. So there's some other data out there in the market. So you could interpret to include just so that you're at least aware of like my overall carbon picture would need to include this.

The wall assemblies-- we went through, it was really important to understand the assumptions. And the help page-- I just took a quick snapshot here of the cladding systems that they've included. And it was really helpful, after having a quick conversation with Jack and Ellis, to go back to the Forma help page and be able to go through and understand exactly what they've got in the assumptions in terms of-- it is different.

And that would be something-- I think they're talking about iterations-- like, as an LCA practitioner, sometimes we want to get into the nitty gritty and actually change the backing material, or change the waterproof layer or something like that. So I don't know if we need to go to that level at these early stage models, but it was helpful to understand the assumptions that are built into the model so that I can have a basis of comparison. And then the last one was also just understanding that it's the 50th percentile results. So as I touched on it, to me, it was like, this might seem lower, but it's because one of the reasons is because it's also best practice.

So we would typically build out our baseline using traditional construction, which might be the more conservative approach. But the Forma tool is using that 50th percentile best practice approach. Pardon me. So it was just useful to see and to understand what's built in.

And I think that's it. Yeah, we're going to-- where do we take it from here? We have a few of these embodied carbon stars in the firm across several of our offices. And we're really excited, I think, to take what we've learned on this project. And we're working on many other projects in Forma right now. So to be able to help educate both our embodied carbon and sustainability teams as well as the design teams on the tool, and hopefully apply it in many more use cases. I'll turn it back to Ellis.

ELLIS HERMAN: Yeah, thanks, Jay. That was really great. So what we really want is to hear from all of you about what this tool accomplishes well and what is left to do. So we have some specific features that we've mentioned on to todo list-- increasing the transparency and customizability of the data behind this analysis, stuff like supporting imported buildings rather than just natively drawn ones. And then more broadly, an understanding that right now, we're here, where we're providing what we think is a really useful, accessible service for the planning stages of design. And overall, we're sort of here with a lot of really useful tools for different personas at different stages of design.

But what we need is this whole project carbon workflow that allows different personas to use their expertise to iteratively consider carbon throughout the entire design process. So we very much encourage you to reach out to us to continue this conversation. Thank you.