.png)
Eldar Sadikov: I believe we're we're sort of entering this golden era for construction technology, and AI is a big element of this.Intro: Built this week, breaking it down. Built this week, we show you how. A fresh idea, a clever tweak you locked in true. Built this week.Sam Nadler: Hey, everyone, and welcome to Built This Week, the podcast where we share what we're building, how we're building it, and what it means for the world of AI and startups. I'm Sam Madler, cofounder here at Rise Labs, and each and every week, I'm joined by my friend, business partner, and cohost, Jordan Netzner. How are doing today, Jordan?Jordan Metzner: Yo, Sam. Happy to be back. Another crazy week in AI. We've seen awesome launches from all the Frontier Labs and, obviously, great guests today, so super excited to get chatting.Sam Nadler: Yeah. I'm really excited to introduce our guest for this week's episode, Elder from Field Materials. How's it going today?Eldar Sadikov: Good to meet you guys. Sam and Jordan doing well. Thank you.Sam Nadler: Awesome. So we will jump into Field Materials. And before we do, however, don't forget to like and subscribe. We have episodes out every Friday. Elder, from tell us a little bit about Field Materials before we jump into the demo.Eldar Sadikov: Yeah. Field Materials AI, we are AI platform to automate material and equipment procurement for contractors, for commercial and and civil contractors that buy lots of materials. We build AI agents that automate all the data entry and a lot of the workflows related to procurement.Sam Nadler: Amazing. Okay. So in our quick conversation, we discussed a few, you know, potential future features that may be relevant to Field Materials AI. I took my stab at at building one of those, this kind of scenario control for different potential price fluctuations and different analyses that that could happen. I may be completely off base here, but let's run through it and tell me kind of what what lands, what doesn't land.It all starts with uploading a BOM, a BOM as I think the industry term. So confirm, upload, boom. We've got all our potential prices here, and you could theoretically go one by one and, you know, manipulate the the price for concrete. Obviously, a very important material based on this project. You know, a 10% increase in concrete could result in a 6% risk of overrun.You know, let's drop that back back down. You could go to, you know, electrical, etcetera, etcetera. I thought these shock mode presets could give us a little bit more to talk about. So if there's a potential tariff shock, boom. It looks like it could have an impact on steel, potentially reinforcement bars, you know, drive a risk of overrun up to 9%.You know, procurement strategy locks structural steel pricing via forward contract immediately. In that case, let's try one more. Maybe there's a port delay. Boom. Uh-oh.That would have a actually, not so bad. You know, a risk of an overrun only 1%. Your your risk exposure here looks to be primarily in plumbing. So your procurement strategy based on this potential shock, you know, scenario would be to diversify your supplier base for plumbing to mitigate concentration risk, yada yada yada. Let's do maybe a market relief and see if that helps.Risk, obviously, nothing. No critical action required. Everything dropped in prices, and, you know, our margin increased. We're flush with cash right now. So, anyway, this is just to kick off the conversation, get started.I have no idea if this is completely off base. Let me know your thoughts.Eldar Sadikov: No. This is this is great. This is actually, you know, reminds me of of a module we have called pricing intelligence where you can kinda see all the prices that you've been paying for materials across different jobs throughout the year and sort of shows you the fluctuations in price. And, you know, we we also have a concept of a BOM where you can upload a bill of materials for the job. But what this does, you know, which is kind of cool is is that you can you can sort of throw in different variables.You know, what happens to to this cost code, let's say electrical or concrete, if, you know, certain scenario happens, and then you can kinda see what what will happen to your total price or to your to your job cost in general. So pretty cool. I would say this definitely has a place and and you know, elements of this you know, could be some of them exist in our platform, and some of them could be ported or you know, could be good enhancements to how how our platform works.Jordan Metzner: Awesome. Let's let's kind of jump into that, if you don't mind, Eldar. Just tell us a little bit about, you know, how did the company start? How did you guys get into the space? And I mean, it seems like AI is so critical to how your company operates, but, you know, did did you guys exist before AI, and how has AI kind of allowed you to to build what you guys are building?Eldar Sadikov: No. That's that's yeah. I mean, we we started in 2022, same year when Chad GPT came out. We started in January, and Chad GPT came out towards the end of the year. The idea for the company I've I've been I've been operating in AI for for a long time, so I I went to Stanford for my grad school, and then I I founded my first company in 2011, which was AI company for retail.This is my second company, and when we founded Field Materials, the basic idea was very simple. The basic idea was, hey, know, commercial contractors buy lots and lots of materials, it's a very manual process, know, you have to you know, get quotes from the suppliers, you have to enter purchase orders, create the purchase orders, you know, a lot of times doing manual data entry. The materials get shipped to the job site or to the warehouse, somebody has to check what has been received, what's still outstanding, what's back ordered. Then the invoices come in, has to verify those invoices, enter those invoices into the account and software, very you know, very verify their validity and make sure that we you received all the materials. A lot of manual work sort of from from, you know, purchasing staff, accounts payable staff, project managers.And the basic idea was that, and can we can we automate some of these workflows? And coincidentally, the same year that we founded the company, AI really made a a a a a a center stage appearance, and we realized in in early twenty twenty three that so much of this could be automated because of the fact that we can read the information from the documents themselves, whether it's a quote document that you get from a supplier or you get a delivery ticket, you know, when when you the material gets shipped to you or the invoice, so much of that information could be read. And then by doing direct tight integrations with the underlying ERPs, we can automate a lot of the workflows and the data entry. And that's sort of like what laid out the the the product and and really shaped the product, and we launched the product in in the 2023, and it really took off, you know, in the past two years we've grown, you know, three x last year, more than, you know, about seven x the previous year, and really hit it, you know, humming on all cylinders.Jordan Metzner: Awesome. And can you tell us a little bit about like I mean, I guess your your customers are maybe large contractors or large builders, or are they kind of project management companies or even material providers? Tell us a little bit about who's using it and and like, you know, again, I mean, there's construction, everything from, you know, building a barbecue in your backyard to, you know, obviously, like the World Trade Center, so everything in between. But, yeah, tell us a little bit about who's using it and kind of, you know, where they found the best use cases for.Eldar Sadikov: Yeah. Our primary customers are, I would say, two two two camps. I would say 80% of our customer base is specialized trade subcontractors, but commercial subcontractors. So I wanna say that this is not your your you know, don't think of it as residential contractors that do your your home renovation, let's say. Think about commercial contractors that build hospitals, data centers, big office buildings, and in that segment, think about specific specialized trade subs like the electrical contractor, mechanical contractor that does all the hvac and plumbing, the concrete contractor, the drywall contractor.So those guys they they they do this massive project where they buy lots and lots of materials, thousands or hundreds of different and that's that's our primary customer base, and they and especially with the the some of the latest trends in the industry where they build in this massive data centers, you know, the size of Manhattan or you know, billions of dollars, they have very very aggressive timelines to build those those those those projects. And they really need desperately need automation tools to to streamline the the whole process of procurement, what what are we buying, when is that shipping, what have we received, have we paid for this, you know, make sure that we're we're actually using the correct price and that we pre negotiate upfront. All those things that that go into into procurement inventory, what what's sitting in our yard, what's sitting in our warehouse, all those pieces, prefabrication, what are we prefabricated off-site, know, what are we delivering onto the site. So all this all this steps and and AI is a big big element of of what we do in in this because we automate so many so much of many of these workflows.Jordan Metzner: Yeah. I mean, I I guess I didn't know the industry that well, but I presume even today that's still done a lot by pen and paper and spreadsheets and clipboards and things like that?Eldar Sadikov: Yeah. Most of this, the way it's done now in the industry, it's a lot of it is is emails, spreadsheets, and manual data entry into the underlying accounting software. So, you know, even if you put it into the accounting software because it has to be done manually, somebody has to actually type every single line item with all the quantities, all the units of measurements into the you know, on premise a lot of times office software, and we automate all of that. So we basically take that work that has to be done by someone entering all this information, we automate the data entry and we automate all the verification steps so that you still have a human in the loop, but you you take that that that busy work or that that that work that everybody hates. Nobody likes data entry if you if you talk to people, or or nobody likes going and verifying is this correct or not.And that's and that's what we do as a company and alleviate a lot of this pain.Jordan Metzner: Awesome. And so tell us kind of the impact. Is the impact kind of like they're able to deliver faster, they're saving money, they're saving time, all of the above. Obviously, these are not like high-tech companies, so them building their own kind of like, you know, vibe coded AI solutions probably not gonna cut it, especially because you mentioned these are large projects over a large amount of years and things like that. But, yeah, just tell us kind of the net impact on implementing the software to the to the bottom line.Eldar Sadikov: Yeah. No. I mean, the primary there are three, I would say, levers, but the primary impact is just overhead costs, just being able to accomplish more with less. What I mean by that is that if if you're a contractor that's doing a a $100,000,000 in volume of projects and you're gonna do 200 next year because you have a big backlog and you have maybe one or two people in your purchasing or accounts payable, now you can you can do the same volume or sorry, twice the volume that you did last year with the same staff because now you've automated so many tasks. So instead of doubling your your head count, now you can actually accomplish the same amount of work sorry, twice the amount of work with the same with with software, with the same staff.Or sometimes we also hear from our customers like, hey, I have my AP person retiring, I need automation because I'm not gonna be able to replace that person, how can I use software in that? So that's sort of an example of of sort of like automating the human labor and and or scaling humans to more to more volume, to more work. The second the second value is is actually about helping you with the the the material costs themselves. So like so much of the margin is bled because you either generate material waste, you double order something, or you you you pay in retail price instead of paying a a pre negotiated rate that you you you went and negotiated with your supplier. A lot of the things that we do around procurement, automation, verification is to ensure that you you pay the correct price with the supplier, you cut down the rogue spending, and and do all of that.And lastly, I would say the last value prop is just, you know, automating the the the payment cycle, the the procure to pay cycle, which is a lot of times, you know, the invoices takes thirty, forty five days to get approved, to review to review them, and it takes a lot of time to bill to pay all of your bills. We cut that time to literally a couple a few days because we can automate so many of the steps, which means that now you can you can you you know, not only get through the payment cycle faster, but you can also get some discounts because of the because paying your paying your invoices early essentially.Jordan Metzner: Oh, that's awesome. And do you think can we just jump in a little bit to kind of the impact some of these frontier models have had? I mean, you know, have have have the impact of kind of the new Opus models or the new Cloud Code models been step function improvements in in what you've been able to develop, not only I guess from your product perspective, but you know, internally for software development and things like that as well.Eldar Sadikov: Yeah. So if you can see on the screen right now, like you have all these different agents that we build in, so we have we have dedicated agents for different document types. So you have you know, there are different quote documents whether it's switchgear quotes, HVAC quotes, you know there's different documents like trucking tickets or asphalt tickets. Underneath it all, we're basically building a dedicated agent that can understand and read those documents and do the data entry for your understand what is what in that document to automate the verification and matching. Because we're using these foundational models, when these foundational models get better every single quarter, every single month, our fundamental ability to read, interpret, and verify the documents is also going up.So and and and that's that's the that's the amazing impact of all these upgrades that we're seeing in the industry that you know, you sort of like, you know, your models could get better overnight literally because simply somebody released a new version and it works so much better. So and that that's sort of one types of agent that, and now we have 20 plus agents that sort of read different types of documents. And we're gonna be introducing other agents very soon beyond just extraction and reading of the documents, but automating a lot of the steps itself in the ERP or or sort of procurement steps that that need to take place. And because we're aligned on the foundational models underneath, obviously our models will get better in that end. And we also ourselves doing a lot of work just constantly refining the models themselves and training those models on thousands of documents that we have, sort of fine tuning them to to operate better and better, and and making a lot of enhancements to to make them smarter.Jordan Metzner: Awesome, super cool. Just give us some maybe just as we kinda wrap up a little bit, but you know, where do you see the construction industry going and, you know, how does how do you see AI having a larger impact on the industry overall?Eldar Sadikov: Yeah. I mean, construction is one of those industries that have been for the long, you know, for a long time has been kind of lagging behind almost like. Like, when you think about technology adoption, like a lot of people like to to to quote this, like the IT spend of construction industry is so much lower. I think it's like a couple of percent compared to like health care or compared to retail, compared to banking where where companies spend a lot more on IT to to to elevate the productivity of their labor. In construction, we haven't seen this.The productivity of labor did not go go up a whole lot in the past couple of decades, and technology spend was not very high as well. But we are seeing that the the wheels are actually turning right now, like, literally in the past, I would say, five to to seven years where more and more construction tech software companies are coming out and construction company and and one of the biggest kickers almost like that happened to the industry was, one was COVID, but more more importantly is the data center boom. I think the data center boom that that we're seeing now is pushing construction industry to completely elevate its game because because this this this this companies like Google and Matter and and Verizon, everybody else that's building the data centers, they come to these construction companies with these aggressive schedules and and big jobs, that construction companies say, hey, we cannot be operating with the old means with you know, pen and paper methods or or you know, data entry or doing those things, and they go and say we need a modern solution, we need a more automated means, the orthophysical AI, you know, adding robots at the construction sites, doing a lot more prefabrication off-site.So those these massive jobs really pushing the construction companies to innovate and and and and change, and and and that's why I believe we're we're sort of entering this golden era for construction technology, and AI is a big is a big is a big element of this component of this.Jordan Metzner: Oh, that's awesome. Okay. Well, honestly, it was really great to hear and just something I, know, don't really think about so much is kind of how AI implements industries that are so as maybe like, you know, more hands on type of industry like construction. So it's really cool to hear that, you know, these kind of more old school industries are starting to adopt AI. And obviously, huge opportunity, you know, in the commercial space.And then as you mentioned, data centers is just all the all the all the rage right now. So super cool and really great learnings. Alright. Shall we jump into the news, Sam?Sam Nadler: Yes. Absolutely. So let me queue up kinda two articles that, to some degree, are telling two opposite sides of of the current kind of story of what a Claude code and what's happening to some software stocks over the last couple weeks. So this came out, I believe, yesterday. Software stocks rebound as Anthropic announces new partnerships.I believe, you know, partnerships with Slack, Intuit, DocuSign, LegalZoom, etcetera. And then another article, SAP users question, value for money affirms AI tools. I think just generally, like, you know, are businesses getting the same value from their historic, you know, SaaS solutions? Jordan, any high level thoughts on, you know, is this a moment in time? Was the, I guess, you know, decrease in in software stocks of weeks justified?Was it just market scare? What are your high level thoughts? Elder, I'll I'll kick it over to you.Eldar Sadikov: My comment on the first news about the Anthropic and and Cloud sort of release releasing this enterprise plug ins is that I think it's a very logical step because I think the first wave of AI has been predominantly just like, you know, get it you know, people would use AI to get advice or or summary or, you know, sort of like like like just, you know, as a chatbot. I think the next wave that we're seeing with OpenClaw and some of the the most recent developments is that you're given hands into the the into the hands of AI and and you ask an AI to do something with your data, to do something, you know, with your with your tools. And I think that's a prime example of that development where, you know, Anthropic is realizing that, you know, this is critical in order to to to to give AI arms or hands. And you know, being able to plug into the enterprise tools that these companies are using is a logical step. You know, Gemini already kinda had had had obviously an advantage because because they're it's a Google tool and they have they have inherent ability to plug into some of the things like Gmail and Calendar.But but this is powerful and it's and and it's it's a it's a it's a longer running trend. On the second news, however, on on SAP, I I think it's almost I think while, you know, the both news are kind of related, I think that on the with with SAP news, think it's it's more of a a different story to me. I think the story is more like, yes, you have a lot of this incumbents and a lot of the enterprise software play players that are trying to adopt AI tools, but the key is not to build a gimmick. The the key is not to just check the box and say, yes, we have an AI tool in our in our in our in our toolbox and now you can use AI with with with this with this old enterprise software. The key is to make it useful.Like at the end of the day, you build an AI software like like us, you have to go from the use case. What are you what can you fundamentally do different for the end user that you cannot do with the current software? And once you have a use case, then you go back and actually build the AI to to to actually accomplish and and and realize that use case. If you do the opposite where you where you simply say, hey, I have a piece of software, let me add AI to it, what's the what's the minimum thing I can do and how can I, you know, you know, launch it as fast as possible, then you end up building a gimmick and and nobody nobody needs that? And I think that, you know, the the the key is to to actually approach it the right way and and build something useful.Jordan Metzner: Yeah. So I I think I'm probably in a pretty similar camp. I think, you know, you talk about these incumbents, whether it's SAP or Salesforce or, you know, some of the other ERP solutions, you know, obviously, they've been terribly slow at implementing AI on top of their platform. So you have this natural behavior happening, you know, pretty organically, which is companies are saying, hey, you know, we're relying on these massive, you know, systems of record, whether that's Workday or SAP or whatnot, you know. And we're hoping that they implement AI because, you know, it will make our lives better, it'll make our job easier, it'll help us manage our company better.But the companies aren't delivering on those promises or certainly not delivering on on those products. And at the same time, you know, while those companies are getting pressure from their investors, their clients are getting pressure. Right? You know, what are you guys doing with AI? Right?And so, you know, over the last probably few months, we've seen a lot of people start to kind of build on top of. Right? Like kind of, hey, just leave it where it is. Let's take out all the data we can, and let's go build on top of it with AI. Now, obviously, in this natural organic behavior, at the same time, obviously, the sales forces and the SAPs are trying to build AI to deliver a valuable product.And there's always and I think this is probably where they're getting hit the most, is probably on new customer acquisition. Right? I mean, who really is a large company today that say, oh, let's sign up for Salesforce today? Like, need to sign up right now, like, right this minute. Right?And I think that's probably where they're really getting squeezed because, you know, if you have a if you have a large enterprise and you're sitting on top of Salesforce as like your central CRM for your company, I mean, your ability to move off of Salesforce quickly is probably pretty difficult. Now, building your own like vibe coded CRM may be possible in a short term or for a small type of company, but if you're a large enterprise, you have thousands of seats or something like that, it's probably not so easy to pivot off of Salesforce, you know, that quickly. It's probably easier to extract data from Salesforce, build some AI on top, build like a mini tool. Right? And I think everybody's hoping, at least some of their clients are hoping, that there's a backfill here where AI comes in and says, hey, don't worry, you don't need to build all these tools on top, you know, we're gonna come and solve your problem for you.Now, if you look at a company like Google, I think, you know, probably six, eight months ago, we were talking about Bard, and how bad their product was, and how they're a loser in the race, yada yada yada. And now, we're pretty much talking about Google kind of in the number three spot from a foundational model perspective, with the probability or optionality to get into the second or first spot. Right? And you know, high-tech companies like Google, we're seeing doing a great job of implementing AI to the point that, you know, what will be the impact on, you know, Google's, you know, work suite, for example, and how will that have a negative impact on on Microsoft's work suite, for example. You know, I saw yesterday that Salesforce announced a deal with Anthropic, but, you know, I mean, Anthropic is both, you know, their friend and their enemy at the same time.Right? And, you know, these companies that own these foundational models, whether that's OpenAI or Anthropic or Google, obviously have this unique competitive advantage where they can go build their own SAP or their own CRM, and, you know, maybe they can even build adapters to do that. And I think I think the news I don't know if we shared the article, but it was something about, you know, figured out a way to write Cobalt code and transfer it into new modern language and look what that impact is on IBM. Right? And so, you know, these kind of more antiquated players, you know, their competition is coming at them faster than they ever thought.I don't think IBM thought they were like in this unique space where who the hell wants to manage, you know, kinda Cobalt software development. But, you know, then, you know, Anthropic let their models run automatically over the weekend and come back with a, you know, Cobalt, you know, transformer, and boom, you know, maybe banks are starting to migrate off of Cobalt because now they can buy this transformer software. So, you know, it is like a moving target, it seems like. You know, I would say that, like, some of these large major companies have an opportunity, but that opportunity is fleeting, especially if they're not able to continue to acquire new customers at the same rate they were historically. You know, and and what would be the impact of that.I think like, there's such an opportunity for Salesforce to just be an AI focused company that ends up becoming a huge player in the space, but it's also looks like it's more possible, if not more probable, that they actually end up losing here because they're not able to implement AI fast enough within their software and their organization.Eldar Sadikov: Well, I would I would maybe add to this that the the it's interesting because I feel like it's also retention play. Like, this this companies, the the big incumbents are launching AI tools so that the, you know, this existing big enterprise customers don't have reason to leave them because they would say, oh, you don't have AI tools to to new players that are trying to to to innovate and build out new, you know, new AI powered CRM or new AI powered ERP or whatever that might be. And but I think it's kind of reminds me a little bit of of the early days of AWS if you think about it because, you know, when when when AWS came up, you know you know, all the big guys did not use cloud computing, and and they used IBM or or something like that. And and, you know but AWS started from the the low end of the market. They said, oh, we're gonna serve startups, and they serve them well, and they moved up the market.And then the the the incumbents like IBM tried to hold on as as long as possible to that to that enterprise segment, and they still hold, you know, some of that segment. But but it's sort of a a very interesting development, very much in line with what was happening to cloud computing a few years back.Jordan Metzner: Yeah. And even as you mentioned, kind of cloud is cloud computing is now having its moment as well as these data centers are getting built, and, you know, you see people buying hardware to to run their own OpenCLAR or whatnot. Right? So, you know, kind of what's old is new and what's new is old, and, you know, we're starting to see the cycles, you know, over and over again.Eldar Sadikov: Oh, I just just wanted to mention that the the one thing I wanted to say about the the SAP news also that as a company that's building AI, it's also a a testament to the fact that AI tools actually are not easy to build. They they actually take a lot of iterations, and the reason why it sometimes they they you know, these these tools get get some of this publicity is because when if you release them too early when they're not polished or they're not fully designed to address the use case that you have, you know, that you know, it's just premature because the tool is not built, and it actually takes many, many iterations to get AI tools right to to address the specific business cases. So I think that is is a testament to how difficult it is still to build, you know, quality AI AI software.Sam Nadler: Haldar, thank you so much for joining. Where can where can people find you? Where can people find Field Materials?Eldar Sadikov: I would say the best way is to go to Field Materials dot com, fieldmaterials.ai. Both URLs work. And if you want to know more about, like, my opinion on some of the things, you can also follow me on LinkedIn. I you know, you can find me on LinkedIn just based on my name and last name. And, yeah, I'm very excited to be a part of this podcast today.Sam Nadler: Awesome. Thank you so much. Great episode. And again, reminder, if you haven't already, please like and subscribe. Episode's out every Friday.Thank you, Jordan. It was a pleasure.Jordan Metzner: Thanks, guys. See you soon. Bye.
.jpg)
