- 2 days ago
AI and automation don’t have to be complex if it’s built on the design standards and rules you already use every day. In this session, a leading builder and an automation expert team up to demonstrate five simple, high-impact use cases to reduce manual work in design and documentation.
Through real examples from dimensioning to wall systems, you’ll see how everyday processes and problems can be streamlined and solved through scalable automation techniques. Walk away with a practical framework for turning your existing standards into automated, easy-to-adopt workflows that improve efficiency, accuracy, and speed to market.
What you’ll learn:
Learn how AI builds on the standards you already know, making it easy to adopt.
Explore five practical and approachable AI use cases that streamline homebuilding design and documentation.
Discover the real-world benefits of AI, including improved efficiency, consistency, and faster, higher-quality results.
Learn the key factors for successfully implementing automation into your workflows.
In partnership with: Higharc
Through real examples from dimensioning to wall systems, you’ll see how everyday processes and problems can be streamlined and solved through scalable automation techniques. Walk away with a practical framework for turning your existing standards into automated, easy-to-adopt workflows that improve efficiency, accuracy, and speed to market.
What you’ll learn:
Learn how AI builds on the standards you already know, making it easy to adopt.
Explore five practical and approachable AI use cases that streamline homebuilding design and documentation.
Discover the real-world benefits of AI, including improved efficiency, consistency, and faster, higher-quality results.
Learn the key factors for successfully implementing automation into your workflows.
In partnership with: Higharc
Category
📚
LearningTranscript
00:00Welcome, everyone. I'm Allison LaForgia, the Managing Editor of the Content Studio at HousingWire, and today's webinar, Simplifying AI for
00:10Home Builders, Five Ways to Streamline Design and Documentation, is produced in partnership with HighArk.
00:17Now, we're going to jump in. Jennifer and Warren have a fantastic outline of things to go through with you
00:23today. I'm going to take a moment to orient you to your screen.
00:26At the top of your screen, you'll see an option for chat, ask a question, resources, and even emoji reactions
00:32at the bottom of your screen.
00:33This is meant to be an interactive conversation. If you have any questions for Jennifer or Warren, drop them in
00:39either the Q&A or in the chat.
00:41We'll be keeping an eye on it throughout today's conversation. A recording of today's webinar will be sent to all
00:47registrants early next week,
00:48so if there's something you want to go back through or watch again or send to anyone that you know,
00:53no worries. You'll be able to do so.
00:54Now, I'm going to pass things over to Jennifer to jump into today's content. Jennifer, the floor is yours.
01:03Hey, everybody. Thanks for joining us today. I just want to say we're going to spend the next 45 minutes
01:08on something really practical.
01:11This is not a demo. This is not a prediction about where AI is headed, but it's really a framework
01:17about how to actually think about this
01:19and what steps to take, because I don't know about you, but I think it's kind of intimidating to think
01:25about how we're going to end up using AI in home building.
01:28So, I'm Jennifer Hoops. I am the director of solutions engineering here at High Arc. I come from home building,
01:35so my parents owned a custom home building company when I was a kid.
01:39I went to school for architecture and urban design. I worked for my family business for a few years.
01:46My dad saw the writing on the wall around about 2007, came down the hall and said, I feel like
01:51I should sell this land, and he did that, which we now know was really brilliant timing.
01:57In that time, I went to work at the big builders. So, I've worked at three of the top 10
02:01in architecture and estimating roles,
02:05but I came to High Arc about six and a half years ago, really looking for something different. I was
02:11kind of frustrated, the tools that I had been given to do what I was responsible for.
02:16And so, I purposefully was trying to find something disruptive, and I certainly did. So, I was lucky enough to
02:23find the founders here at High Arc.
02:27And ever since, we've been working on a new and different way to solve some of these problems that we
02:31face in the industry.
02:33So, I'm honored to be joined today by Warren from Ashlar Homes. Most importantly, Warren is, he's not an analyst
02:40or a vendor.
02:41He's a real-world builder, just like I hope lots of you are, who's living in this space just like
02:47the rest of us.
02:48So, Ashlar is a High Arc customer, and Warren can certainly speak to what that looks like.
02:53So, Warren, maybe introduce yourself and your company for everybody.
02:58You bet. Thanks for having me. I'm Warren Brenner.
03:01I am COO of Ashlar Homes. We are a privately owned production home builder in Kansas City.
03:08We build single-family duplexes, and later this year, we're going to dip our toe into townhomes and see how
03:14that goes for us.
03:16As Jen mentioned, I am a higher customer. I would say I come from a tech background, so I'm definitely
03:22tech-leaning.
03:23I'd like to think or hope that we're midway through our journey, so we've tried some things that have worked
03:28and haven't worked.
03:29But I guess my job is to tackle all those fun operational challenges that come with growth.
03:36That's great. Awesome. Okay, I'm looking forward to hearing all about what you're working on, Warren, and sharing everything.
03:44Okay, so I want to start with a little bit of a poll here, right?
03:52I'd love to hear from the folks on the call. What types of AI tools has your team experimented with?
03:59This is going to help ground us in understanding a little bit about where you guys are today.
04:05So if you want to go ahead and flip through that, we will tell us a little bit about what
04:11you're working on.
04:12We'd love to hear from you.
04:15So as you guys are filling that out, I'd like to also talk a little bit about a framework here,
04:21right?
04:22So what I want to describe is going to really set the stage for everything else that you're going to
04:28hear today.
04:29I want to reference two very well-known architects from history.
04:34One is Corbusier, and the other is Christopher Alexander.
04:38So the first, Corbusier teaches us in architecture school that we should treat a building like a machine.
04:44So we should define the problem, standardize the solution, optimize for efficiency, and it should hold.
04:51I don't know about you, but that sounds really hard to do in home building.
04:55Predict, define the problem, predict what's going to happen, and define one solution that's going to fit.
05:03Christopher Alexander looked at this problem very differently, right?
05:06Real environments don't repeat themselves exactly, but they do repeat problems.
05:12And I hope that that's what we get to talk about today, the problems that we find that are repeating
05:17in our business.
05:18So Alexander would tell us, instead of fixed solutions, you design patterns.
05:24So relationships that adapt as these problems resurface.
05:28So why does this matter?
05:30Construction really behaves much more like patterns than it does a machine.
05:35We all understand that we are not building houses in factories with isolated conditions where we control all of the
05:43factors.
05:43We're building homes in real world on real lots with real dirt and with human hands.
05:49And that means that we need to manage these variations much more efficiently.
05:55So before we get into examples, I want to just name something very specific.
06:01These problems that we face, you all are tackling them today, right?
06:05So Warren, you're a great example of this.
06:07Your team is doing this work, right?
06:10The problem isn't necessarily that we can't do the work.
06:13It's just the pace is really overwhelming.
06:16So plans need to be done.
06:18Jurisdictions need changes.
06:20Options need to proliferate across our business.
06:25It's just a lot of work, right?
06:27And so the workload might be manageable, but the speed of this change and what we're dealing with is not.
06:34So it's not that your team is failing.
06:37It's that the tools that we have built are really designed for a more stable system.
06:44They're designed for a one way of working, and that's really not what we see today.
06:51So Warren, I don't know if you have any comments about something like that, but any thoughts about like, is
06:58it something, is that right, right?
07:00That we technically can do the work, it's just really a lot of heavy burden on our teams.
07:09I definitely think you're right, and I think, I mean, it goes in every step of the process in home
07:14building.
07:14So from permitting to plans to purchase orders and all the way out through the build process, I think, you
07:23know, one option is to add more people, throw more people at the problem, increase your costs, manage overhead, stuff
07:29like that.
07:30Obviously, if you're trying to stay lean, build a viable business that has good margins, that's not, in my opinion,
07:39the right way to go.
07:40Plus, the more people, I would say, maybe tends to be the more problems.
07:45So for us, a good example of that is purchase orders.
07:48You know, we were starting to do more volume, so we thought, okay, we got a spike, so let's cross
07:54-train another team member and let's teach them that.
07:58And then when the spike goes away, then we'll be all good.
08:00So, one, like, not having our rules codified for how we do purchase orders, you know, there's knowledge gap transfer
08:09in that transfer.
08:11So we did get a second person, but they proliferated some, you know, errors out there because they didn't know
08:15what they were doing, so we created more problems for ourselves.
08:17So in that case, adding more people certainly didn't solve the problem, so we just needed to codify, you know,
08:22the process and do that across everything we're trying to do.
08:25So, right.
08:27I love that because basically what you're saying is that more people, if there's chaos in the business, if you
08:34don't have your operations structured and your data structured,
08:39more people actually just are dealing with the same chaos, it's just spread thinner, it's spread across more individuals, right?
08:47And so what we need to address in order to kind of move the industry forward is really setting up
08:55a better structure for our data and the way that we run our business.
09:01So I do want to point out that I think this is exactly where AI patterns go sideways, to be
09:09honest, right?
09:10So people will say something like, well, you know, can't AI just do it faster?
09:16Well, my question is faster what, right?
09:18Like, what are you trying to go faster with, you know?
09:21And I think that's where we want to highlight these examples today, where patterns emerge in our business, they're really
09:30opportunities for us to automate.
09:33And outside of talking about individual AI solutions, which we're going to get into as well, but outside of that,
09:39I think it's really important for everybody on the call
09:42to recognize that for you to use this new technology, for you to take that first step towards a AI
09:50and automation that we're feeling this need to do,
09:54you have to recognize what it is that you need to solve within your own business.
09:57And Warren, that was a great example that you gave us of your POs, of something that, you know, an
10:04opportunity for you, to say the least.
10:06Yeah, for sure.
10:08Okay, so I want to just kind of jump in and be very specific about this around why patterns are
10:15useful, right?
10:16So I don't know how I wound up as this, you know, liaison of tech in home building, but here
10:24I find myself having the unique opportunity of talking with all of these really great people who understand how AI
10:30works.
10:31And in reality, it's very much about patterns, right?
10:36So accurate predictions using historical data is really what all of these large language models are doing.
10:43And so that's what we need to look for in our business, our patterns, right?
10:49Once we identify the patterns, we are able to start to recognize what that means for us as unique home
10:57builders.
10:57Okay, so let's look at some really common examples of this.
11:02So rework, rules, and we're going to get into very specific examples in just a second, variables, costs, all of
11:12these things should be something that we start to recognize as a pattern in home building, okay?
11:19So let's jump into some specific examples.
11:25Homes are made up of common parts, but we keep doing this work over and over again, right?
11:31So across your plan library, if you're looking at CAD files or Revit files, if you're looking at estimates, if
11:38you're looking at floor plan, marketing collateral, all of these individual assets contain the same types of data, right?
11:47And so if we look at how we manage that data, it starts to unlock a lot for us, okay?
11:55So let's get real specific.
11:58Let's be very practical.
11:59So we'll talk about some real-world examples here.
12:03This is where I kind of want to intro a little bit of a macro problem that we're dealing with
12:09and how it surfaces in the real world.
12:12So oftentimes, we're looking at smaller communities than we did 10, 15 years ago, right?
12:19I was just at a land event in Charleston that brought people together to talk about this problem.
12:26Like land light versus land heavy means that a lot of times we have more communities.
12:31Those communities all have different lot sizes, and we have to meet their requirements for each one, right?
12:38And so what we're seeing is if I could change my garage orientation, if I could change my height and
12:44width, if I could change my overall size of my house, if I could proactively identify these types of things
12:53as a pattern, we can start to use technology like HighArk to manage those patterns and really make a system
13:03that captures automation.
13:05So Warren, I'd love to hear how this shows up in your world at Ashlar, if you're feeling this kind
13:11of pressure to use plans and kind of automate some of the variables that you can see happening over and
13:21over in your business.
13:23Absolutely.
13:24And I could say when we were doing, I don't know, 50 homes or 60 homes, probably the same problems
13:30were there, but they were more manageable because, you know, you could do it manually,
13:34or you could have one or two people that knew the product in and out and could just go do
13:38the work.
13:39But if you're going to build 200 homes or 500 homes across 10 communities and you don't want to proliferate
13:45plans and plan sets,
13:48like we use the data, for example, for monotony rules and for lot fits.
13:53So we use the lot list in HighArk.
13:57We put all of our dimensions for the build width, build depth, as well as garage handling, all of that
14:03good stuff.
14:04And then we'll programmatically go codify the rules on what should fit and shouldn't fit on a lot.
14:11And so then we'll run through there.
14:13Monotony rules are the same way.
14:15Those are a piece of data for us.
14:16So a monotony rule to us is if you're building on a perfectly rectangle street, you're the house in the
14:21middle, the five houses around,
14:22you can't have the same home plan, I'm sorry, the same home plan and elevation, and they can't have the
14:27same exterior palette.
14:29So for a human to go through all of your communities every day, every time a house sells,
14:35and put the right foundation type that that lot allows or put only the plans that that lot allows, it's
14:41just not scalable.
14:42So we codified our rules, we use the data on the lot list, and then every night we run scripts
14:49that basically will disable automatically plan options and plans and or plan options that don't fit on a lot.
14:58So our salespeople never are presented with the opportunity to sell a slab home on a walkout lot.
15:04They can't make the mistake of selling the wrong exterior palette, which breaks the monotony rules.
15:10And ultimately, that should end up in one, a much better operation, but two, I mean, a much better customer
15:16experience,
15:17because you're not having to backtrack and be like, oh, I'm so sorry, I sold you something that we can't
15:21build.
15:21So yeah.
15:23Yeah, that's a great example.
15:25Yeah, no, that's never happened in my career ever.
15:28Right, exactly.
15:30But no, that's, that's great to hear.
15:32I'm happy to know that that's working well for you guys.
15:35And I think too, the way that I think about it is, I can, I can also very, I can
15:41move as a director of architecture in my past, right?
15:44I can start to move my business towards a more strategic orientation than a reactive one, right?
15:51Because this isn't, this isn't always surprising to me, right?
15:55Like, if, if someone like you would come to me and say, hey, we have a need to change our
16:00plate height, to change our garage orientation, to change this house from 48 foot wide to 49 foot wide.
16:09I'm not surprised by that, right?
16:11Like, these are things that in our world are, are, are things that we've had to manage reactively.
16:17And so what, what hierarchy allows you to do is be a little bit more strategic because the plans are
16:23actually data models, right?
16:25And so it's very, it's very easy for us to take something like, oh, a plate height and, and simply
16:30change the direction to our tool for what that needs to be.
16:35And so the reason that, that Warren's able to use this tech to kind of, you know, make sure that
16:40his salespeople are selling what they're supposed to sell is because behind the scenes, we're able to set up all
16:48of these kinds of conditions and circumstances as part of a centralized data model, and then use it across the
16:56library in the form of plans.
17:00All right. I think that another really, you know, good example of this is, is jurisdictional controls.
17:06And I think we started to get into that a little bit, but, you know, I think that sometimes we
17:11think of jurisdictions as affecting design, but in reality, they're not, they're not doing that.
17:17They are introducing rules and we have to modify our designs to meet those rules.
17:22So this is something else that I think that we can, we can start to quantify as a pattern in
17:29our business, capture some of what these things are and, and proactively build a tool that lets us address those.
17:37So let's, let's go through a couple of examples.
17:39So like, I, I know for sure that probably 90% of the time that I was going through an
17:49ARB review, they're going to ask me to change some sort of masonry percentage, shutter style, whatever it is on
17:57the outside of the house.
17:58These are things that we can use our tool to automate the production of, right?
18:05So we can, um, very quickly manipulate at a community level, the, um, controls and, and really what that allows
18:15you to do is move much quicker through creating a plan that exists, a design that exists that meets the
18:24rules for a given location.
18:27So Warren, talk to us.
18:29Are you seeing this across your markets, some level of predictability in, in these requirements?
18:34Uh, I would say definitely.
18:36And for us, like the best example is like energy codes.
18:40So we build in a lot of different municipalities.
18:42So Casey Mo is the most challenging.
18:46They've adopted the most strict energy requirements.
18:50And so normally we would have to create separate plans for each community that have the different code requirements.
18:57But now in higher, we basically have a single plan and we put all the options into that plan and
19:04we bundle that up into a single energy code package, which we make available in Kansas City, Missouri.
19:09And so it get, and you, we mandate that anytime you build a house, that's automatically checked.
19:14It's not optional.
19:15It comes with the house, but those options exist in hierarch and we turn them on as a la carte
19:20options for anybody who wants to upgrade their AC or if they want insulation in their attic.
19:26Um, but we can do all that with a single plan just by configuring that.
19:29And so for us, that's step one, but you know, like people, people change that those codes are going to
19:35change.
19:36They're going to vote something else in there.
19:37So our roadmap is to actually use AI to ingest those, those local municipality codes and, and drive like automatic
19:47updates of those plans, obviously for a human review before you push it out there.
19:51Um, right now our BIM managers are managing that through conditioning and hierarch, but I do think there's opportunity to
19:57use AI to, to be even more proactive and, and rolling those out.
20:01I fully agree.
20:03And I love the idea that because your plans that exist in hierarch are actual data now, right?
20:11What you're going to be able to do is use AI to start to even query those things.
20:16Like a really specific example, um, that I, I was going to going through with a customer a couple of
20:23weeks ago is, well, could I use hierarchs AI to ask it all to identify all of the opportunities where
20:31I have a window within 18 inches of a door?
20:35Now you and I know they were actually looking to make sure that all of their tempered windows were correct.
20:40And the answer is yes.
20:42Right.
20:42Like that's absolutely, I'm with you, Warren.
20:45Like that's a really good, good goal and a good target for us to, to use AI.
20:51Now that your, your plans exist as data, we should be able to say, all right, well, across my plan
20:56library, I need to holistic, holistically evaluate all of the circumstances in which I'll need different insulation or different material
21:06or something like that.
21:07Right.
21:07And that's the type of, um, that's the type of, uh, data that's just not, it's not available from a
21:15plan set, if you will.
21:18Yeah.
21:19And even as awesome as hierarch is like our, our bin managers, they're creating the plans and they're human.
21:25So human error occurs, but you could use AI to say, look at this plan set, review this plan set
21:32against the code for that municipality.
21:34And tell me before I said, my, my plans, if it's going to pass or fail, uh, and catch, you
21:40know, just shorten your permitting process.
21:42Right.
21:43Yeah.
21:43And I just, I think the difference here is that, you know, if you were to put a, a plan
21:49set into, uh, a commercially available AI tool that exists out in the world, um, that plan set contains things
21:58like dashed lines.
21:59Right.
22:00You and I, and probably a lot of the people on this call know what a dashed line means in
22:05a plan set, given our experience in our context, right?
22:08A dashed line can mean a case opening.
22:10It can mean a, uh, a closet shelf.
22:13It can be a cabinet, can be a ceiling above.
22:15There's a lot of things that a dashed line can be on a static plan set produced by AutoCAD or
22:22Revit.
22:23But what, what's going to enable Warren to, to use AI to do these things that he's describing is that
22:30while we have decided that we're going to use the dashed line to show those things on your plan sets
22:35behind the scenes,
22:37the actual hierarchic data knows, is it a case opening?
22:42Is it a wall below?
22:44Is it a closet shelf?
22:45It's not just a dashed line to us.
22:48Right.
22:48And so when we're, we're starting to query that data, we're starting to use this large language model technology, it
22:55becomes more accessible and we're able to understand better the information that lives inside of our plans.
23:04Right.
23:04But it's pretty exciting.
23:08Okay.
23:09Um, let's talk, let's talk about another example.
23:12And this, this really is, is something that, um, I think is near and dear to my heart.
23:17Um, I, I think that what we have been doing as well is we have been, you know, a lot
23:24of times in my career, I felt like, all right, well, um, you know, I'm going to draw this plan.
23:29I've got a lot of things that I do over and over again, like electrical plans.
23:33I just drew them the same every time.
23:35How many times did I draw a linen closet?
23:37How many times was the linen closet different?
23:40Almost never.
23:41Right.
23:42I can very, very clearly articulate in my business that a linen closet is four 16 inch deep shelves.
23:51It has a light only if the closet's deeper than two feet.
23:54Right.
23:55Like I can very quickly say that.
23:56But what's been happening in the industry is that we've been drawing that, estimating that, representing that on brochures over
24:05and over again.
24:06So back to the framework that we started the call with, that's a pattern that's emerging.
24:10Right.
24:11That, that I can offload to a technology tool to make me a more efficient person.
24:18I can spend my time on design versus recreating a linen closet over and over again.
24:26So Warren, have you guys found, um, things like that in your business that you can kind of offload to
24:32the tech to help make you a little bit more efficient?
24:35We, we definitely, when we were first implementing it, took the opportunity to, we had very similar plans that we
24:42were able to consolidate down and, you know, standardize on, you know, certain, I don't know, plate heights, all that
24:49good stuff.
24:50Um, and so definitely we had the opportunity there.
24:53Exactly what you were talking about was talking with Sean, our owner this morning.
24:56I mean, we're now at a point where we, we've got an implementer, we're out, uh, you know, in production
25:02and we can start to like, like be intentional about our design.
25:06We, we have like 10 plans and things that I won't say customers don't care about, but maybe their lower,
25:13lower priority for a customer would be like a hall bath or a basement bath.
25:17There's no reason that across our 10 plans that a hall bath should not be a 36 inch vanity with
25:23a pedal sink or, or whatever, a standard set of things to whenever we go to design our next plan,
25:29we say, great, let's drop in that, that hall bath.
25:32And it has all that, all the exact same things, which trickles down from a purchasing perspective, because then you
25:39can slim down your, you know, your purchasing catalog and everybody just knows when you order a hall bath, exactly
25:45what's in that PO and you're going to have fewer mistakes there.
25:48So we are not there yet candidly, but that's where we're going.
25:52Yeah.
25:52It takes it.
25:54I mean, it does take a while to get your head around this like mindset shift, doesn't it?
25:58Right.
25:58Like, like to figure out what these things are and how they benefit you.
26:03Yeah.
26:03There's, there's, it's a natural evolution over time.
26:06Once you start to recognize these patterns.
26:09And that's, that's a good thing that you can be able, you can come back in and, and decide, well,
26:16I'm going to automate my bathroom.
26:18Right.
26:18I'm going to build a layout here across my plans and the tool will, will support you doing that.
26:24So it's great to grow into it and, and start to, to find the individual value for your, for your
26:30team.
26:31For sure.
26:33Okay.
26:34Let's explore a couple more examples.
26:36This is another one that, you know, is just something that the industry has long struggled with.
26:42Right.
26:42Um, how many times have I been asked, well, what if I decided to go back to dimensional lumber instead
26:50of floor joists, because they're, um, they're requiring a, uh, a code update for fire prevention.
26:57Right.
26:58Here I am the plans.
26:59I'm the keeper of the plans and the estimates.
27:01And I'm thinking, man, you, that's going to change everything.
27:04Right.
27:04Because you're going to change the dimension of the floor system.
27:08I got to update my stair.
27:09I got to do all of these things.
27:10And so it became a, a barrier to you, someone like you, Warren, managing the cost of your homes, because
27:18you might, you might see a better, more efficient building system situation.
27:22Doesn't affect the design.
27:23Doesn't affect the usability of the home.
27:27But because our systems have been so disconnected for so long, we've been kind of locked in to the original
27:34way that we decided to do things.
27:36So, again, I'm going to sound a little bit like a broken record.
27:39This is a pattern, right?
27:41This is a pattern emerging that we need to be able to control the way that we build houses in
27:47order to be the most efficient.
27:49Right.
27:49And so it's the same kind of thing.
27:51This is where we can look to automation.
27:54This is where we can, we can say, all right, I'm going to identify the building systems that I need
28:00to be able to manage independently of any one plan and really become much more strategic in our, in our
28:08operation.
28:08Does that resonate with you, Warren?
28:11Unfortunately, yes.
28:13I mean, fortunately, but unfortunately, I mean, the reality and, you know, historically, right, before, before Hirec and just in
28:21this business, change is inevitable.
28:23And how well or how capable you are of managing that change is probably going to predict your success because
28:29there's always a new design system or maybe there's a new idea or maybe there's a new code change.
28:34The reality is you've got it in your plan right.
28:37And then you sell 20 houses over some period of time.
28:39And you've got 20 houses with that plan at various stages of construction.
28:43And then you have a change.
28:45And then so some of those, some of those need the updated plan.
28:49Some of it's way too far to get the updated plan.
28:51But having a, a, a systematized way to go say, great, we've got 20 of these plans out here, three
28:58are at permit or at final walk, whatever, but actually being able to scan and should be changed.
29:06And then changing them is as easy as like making the change in Hirec and releasing it for that update.
29:12And then obviously communication with your field, getting the right stuff.
29:16But, you know, having a tool that allows that is, I mean, I don't know how you, I know how
29:21we didn't do it well beforehand, but now at least there's a way you can do it.
29:26Yeah, that's right.
29:27And I, I love these examples too, because I, like, I remember, I was at a frame walk once in
29:33this, in this quiet, quiet gentleman that was at this like frame walk.
29:37We were there for value engineering, right?
29:39We were trying to come up with ideas on how to get costs out of the house.
29:42And he came up to me and just tapped me on the shoulder and said, hey, I've, I've had, I've
29:46always wondered why you don't move your second floor windows up a couple of inches so that the freeze board
29:54comes down far enough that you don't have to put brick above the window.
29:59Right.
29:59And I kind of look at him like, whoa, that's so simple and brilliant.
30:03Like, why didn't I think of that?
30:05Right.
30:05You just eliminated the lintel, the brick and all of these things.
30:08I go back to my office and I realize I got to put, like, this is going to cost me
30:13hundreds of hours to update this in drawings.
30:16And so this is a cost savings opportunity that I can't capture because my system is not built for, for
30:26automation.
30:27It's not built for agility.
30:28And so that example in high arc is a single input where we would change the window head height that
30:34would, that would allow an operator to capture that value.
30:40Right.
30:40And I, I, I think that, you know, as we were saying, you're, you're growing into this concept and how
30:46this is going to affect you in your, in your role at Ashler, of course.
30:51And I think continue to look for those opportunities as exciting and have access to, to really control those costs
30:58is, is, is very cool.
31:00So.
31:00Yep.
31:01Absolutely.
31:02Okay.
31:03Um, I think our last example, and then we want to get into, um, a little bit more, uh, kind
31:09of free, free form talk about what you're working on more.
31:12And our last example, I want to kind of, um, plant with the, the, the, the, uh, audience here, um,
31:19around these patterns is designing costs without feedback.
31:23Right.
31:23So I'm sorry, I said that back to designing without cost feedback.
31:28Right.
31:28So here we go, designing a house or designing an elevation.
31:32This is not unknown, right?
31:34This, this shouldn't be unknown.
31:36I remember sitting in my office thinking, okay, well, they told me to take this elevation from this community to
31:42a new place.
31:43And it's supposed to be cheaper than the one that I, I started with.
31:48Well, I don't know what I'm doing.
31:50I don't know if I change this from, uh, shake siding to vertical, if it's actually as cheap as I
31:58need it to be.
32:00I should be, I should know that like the company knows how much that siding costs.
32:04Right.
32:04So there should be some sort of way that we can connect these things.
32:09Right.
32:09And so this is a, this is a journey that a lot of our customers are on.
32:13And, and to be, to be honest with you, it's a new way of thinking about connecting costs and it
32:18takes a little bit longer, but setting this system up to automate the connection of cost to design is also
32:27really, um, really a great opportunity for us to bring automation into, into home building.
32:32And so, um, you know, what that means is that, that it may not, it may not be a direct
32:39house cost decision, or I mean, a direct house cost result, but what it means is that I spend less
32:46back and forth with my team.
32:48I can produce a better initial, uh, home because I'm more in touch with the costs that are real to
32:58my organization.
32:59What would something like that mean to you guys, Warren, if you were able to kind of connect these costs
33:04early in your process and, and make better informed decisions?
33:09Uh, I think, I think it's, I think it's everything to be honest with you, because Sean, again, our owner's
33:14been doing this forever.
33:15We just launched a new community.
33:17We put out two new, two new townhomes that we've never built before.
33:21And when you're, you know, doing your preliminary, should we buy this plot of land?
33:26Can we make money with houses on this?
33:29You pencil it backwards and you're like, oh yeah, if we can sell them for $3.95 or $2.95,
33:33no problem.
33:35And, and, you know, he's going on gut that says, yes, I've done this before.
33:39I'm pretty sure based on the materials and the size of the house and all that good stuff, I feel
33:43like we can do it.
33:44The reality is you go do all that stuff and you put your plans out for bid, it comes back,
33:50you do all that good stuff.
33:51And not until like it's way too late, maybe you find out, oh my gosh, I didn't realize that this
33:57was a requirement and it's going to shoot my price up $10,000 or what, you know, if you had
34:02the ability to accelerate the confidence in, in, in what that price or what the range of that price will
34:08be like months in advance or ideally even before you ever buy a piece of ground.
34:12Like that would be indivisible to, you know, somebody who's laying out the capital there.
34:17I completely agree.
34:18And it's just, it's such a common example, to be honest, right?
34:22And I think we've accepted it for a long time.
34:25And it's, it's really exciting to, to start to understand that maybe we can make better informed decisions using this
34:32tech, right?
34:33Um, and I think this is just the start we can, we can, we can use AI to pull costs
34:40from an area that we're not even in yet.
34:43Right.
34:43And relate it back to, um, to our, to our data model, if you will.
34:48And so, um, I'm looking forward to the next, um, you know, let's call it a year.
34:53I mean, a lot happens with AI in a year.
34:55I'm looking forward to the next, um, you know, group of, of, uh, of stories that folks like you are
35:03able to, to kind of uncover using, using the hierarchical data and AI to really drive, drive this sort of
35:09efficiency into your business.
35:11It's going to be an exciting time for us.
35:13A lot of work.
35:14Awesome.
35:14Okay.
35:16I hope so.
35:17I do.
35:17I agree.
35:18Um, so I'd love to do another poll here, folks.
35:21Um, if you could, um, let us know, are these examples hitting home for you?
35:27Are you, um, which one do you think makes the most sense for your business?
35:31We'd love to hear from you guys, um, around if you, if you see these patterns emerging, um, in, in
35:38your business and, and how it might relate.
35:41So we'll, we'll take a look at that after you've had a chance to, uh, to give us a little
35:45bit of feedback.
35:45But, um, I'd really love to understand, um, what this all means, right?
35:53So what do all these examples share, right?
35:56So really what has happened in every example that we just went through, um, humans dealt with this problem, right?
36:04Like I said at the beginning, we're not incapable of managing these, these situations.
36:10But what it does is it takes us a lot of, of time.
36:14It takes us a lot of human will, human effort.
36:17We make mistakes.
36:18And as we are, are in this environment where AI is, is just absolutely everywhere.
36:25It feels like, okay, I now know if I look at these examples in my business and what's the most
36:31impactful to me,
36:33I should be able to start to identify what type of AI tool, what situation makes the most sense for
36:42me as a builder to, to capture some of these things.
36:46Just broadly saying I'm going to use AI is not necessarily the best first step.
36:54The best first step is identifying these patterns, what impacts you the most, and ultimately what you want to be
37:03able to solve, right?
37:06And so this is, this is where I think, um, for us in particular here at High Arc, uh, this
37:13really makes a lot of, of, uh, impact, right?
37:16We, we, we start talking with builders like you, Warren, about what is the, what is the opportunity that, that
37:23best presents itself.
37:25And we need to, we need to move the, the, the, uh, we need to move past this old world
37:32way, this old way of doing things,
37:34which is these static assets into a world where this data exists in a recognizable way so that we can,
37:43can really start to use it.
37:45And so with that, I'd really love to, um, hear again, I've got a, one more poll for you all.
37:51Um, what is your, your team done in the last 12 months, um, that you, you have, have tried to
37:59address some of these patterns.
38:00And I think that will help you understand if you can kind of answer these things, it will help you
38:06understand where you have the most opportunity.
38:10Okay.
38:12All right.
38:13So, um, I think with this, Warren, I'd love to hear a little bit more like openly from you.
38:20Um, I know you guys are kind of deep in it, right?
38:23So we've talked about this concept of the framework of why we should be thinking about patterns and, and how
38:29we would use AI.
38:31How is this really manifesting in your day to day?
38:34So how, how, tell us about some of the things that you guys are working on.
38:38Uh, I would say we have a couple of, uh, a couple of like things that we would try to
38:46achieve.
38:46One is to minimize variance.
38:48Uh, I think variance is the thief of joy.
38:52I think I tell people, um, the fewer mistakes you make, the less money that it costs you, the less
38:58time it takes, the happier employees are.
39:00So everything we're doing is really trying to, to drive variance out of our process and our systems, whether it's
39:06dirty data, whether it's inefficient processes or not connected.
39:09We want to do everything we can to predictably spec a home, build to that spec, collect payment.
39:17Um, the other thing is I, we, we want to be a lean organization.
39:21So, um, you know, we want to automate the mundane and let our good people go add higher value, um,
39:28to, you know, solving problems that are like value add problems.
39:32Not like putting out fires because we've set the house on fire ourselves, so if we can do those things
39:37or use AI to go do those things, we're looking to go do it.
39:41And I think you said it like AI is not a silver bullet.
39:44It's like, it's only going to amplify your chaos.
39:46If you've already got chaos or dirty data.
39:48Um, but if you can codify your, you know, if you can get your data right, codify your rules, then
39:55you can put AI on that.
39:56And for a small to medium sized builder, I think it gives you the opportunity to, you know, enjoy some
40:01of the efficiencies that, you know, a national builder has, but, you know, at a much lower overhead costs and,
40:07you know, you can advance your maturity a lot faster.
40:10So that's like philosophically what we're trying to do, um, in a practical way.
40:16Like we're, we're talking a lot about, you know, our plans today, but like across all, all of, you know,
40:23Ashler, like all of our departments, HR is using AI for, you know, creating and, and refining job descriptions that
40:30when we're hiring for a new person, just accelerate like things we've never had before.
40:34So we need to go create those job descriptions and then inevitably when we get, you know, a hundred applicants
40:40back or 50 applicants back, we feed those into AI along with the job description that we put out there
40:46and say, give me the top 10 candidates.
40:47And so a human that would have to sift through 50 resumes now can go in there and pick out
40:52the 10 that we want to talk to.
40:53So the time to onboard that person or to find the right person, again, increasing ideally the probability that we're
41:00going to get the right person.
41:01Um, and then also using AI then to go create the training material that we would to onboard that person.
41:08Probably the only, you know, midsize builder out there who doesn't have a deep library of documentation, whether it's training
41:14materials or scopes of work or stuff like that.
41:17Like we're really using it to kind of accelerate, um, you know, creating that material that we can then use
41:22to train people and onboard people.
41:24Um, permitting is another thing.
41:28That's a great use case.
41:29I just wanted to add, I think that's a great use case because, you know, as, as we are moving
41:35towards this world, it does change the type of people that we, we need to hire, doesn't it?
41:40Right.
41:40Like I need to, I, I, I think do the same thing that, you know, I talked to Claude a
41:45little bit about, you know, who's the ideal person that can do, uh, the types of things that we're talking
41:51about.
41:52Right. So even if you were to, you know, prompt, um, you know, your, your, uh, your tool to say,
41:58you know, Hey, these are the, these are the types of problems I'm trying to solve in my business.
42:03How can I interview for, uh, creativity and concept understanding and things like that?
42:08So I really love that example, Warren.
42:10That's a good one.
42:11What else do you guys work with?
42:13Um, well, and along that vein, I guess, purchase orders is like an easy hot button for me because I
42:18used to do them and they're, oof, that's how I learned construction.
42:21But like in many cases, and for us, for a long time, it requires a person who knows deeply our
42:29product and in every plan and every community, what trade, you know, how they get PO differently, all this good
42:36stuff.
42:36Like you, that person should not have to have the level of knowledge of your product and your operation.
42:43You know, it should, your system should have all that codified so that the POs are created automatically.
42:48They give it a sanity check from a human perspective and it pushes out there.
42:52So that per, so you can hire a much less expensive resource to do your, your purchasing.
42:58Um, and you know, if you've got a, you know, somebody that's a high achiever who can go put their
43:02big brain on solving customer experience or something cool.
43:05So, um, other things we're doing.
43:08Um, so hierarch is one of the systems that, you know, that, that we have, and I'm sure other builders
43:16out there, they've probably got an ERP.
43:17They probably got, you know, maybe they're using builder trend or a thousand spreadsheets, but, um, connecting the data across
43:25hierarch to the ERP, to your builder trend and your cycle times and all that good stuff.
43:31So that you can really drive insights in your business is today it's very manual.
43:36And unless you're using, you know, some monolithic system, um, which probably most midsize builders are not, um, it's really
43:42hard to, because there's so much human error in data entry or, you know, just entering a job and builder
43:48trend, um, you know, it, it, there's fraught with error.
43:52So you can, we're using AI to basically connect the data, aggregate the data across all of those systems, and
43:58then go look and say, okay, great.
43:59I've released a spec and it is, you know, at, um, you know, it's, it's a, you know, framing stage,
44:06but it's still in the pre-planning stage and hierarch.
44:09So I need to go in there and I need to move that thing to the right stage.
44:12So when I run my report and see what options are getting bought, I'm only looking at options for houses
44:17that, you know, have been purchased or, are specs.
44:20So you can really do a much better deep data analysis and get a sense of what your, your company,
44:25how it's performing, um, what options are selling and not selling in certain communities versus not.
44:31So when we release our specs, which is another thing we're using it for, you know, you can look at
44:36your hierarch data, you can look at your, you know, your, your sales data to say, what have I sold
44:41in the last six months?
44:42What options were in there?
44:43And then you can also pull in MLS data to look at, you know, your market to say what price
44:49point and what communities and what types of plans are selling.
44:52And then based on, you know, all those various data points say, great, I should be releasing six specs a
44:58week to hit my even flow targets, make a prediction, tell me what I should be releasing based off of
45:04what selling, based on, you know, what we have to offer based on our monotony rules so that I'm not
45:09releasing something that, you know, doesn't exist.
45:11So it's really, it's, it's definitely not a hundred percent dialed in right now, but it's, it's shaving so much
45:17time off of, again, going back to predictability, trying to make sure that you're releasing the right product at the
45:23right time at the right price so that it moves predictably.
45:27Um, and that's a big thing we're working on.
45:30Yeah.
45:31I really like that.
45:32And, and this is, this is one of the reasons, well, there's many reasons we love being a partner with
45:37you, Warren and your company, but, um, we love to hear those types of examples.
45:41So that we can continue to build out, um, you know, easier dashboards for you to, to connect to.
45:49And, and that's how we, we really prioritize some of the work that we do at Highark is, uh, making
45:55sure that we're, we're listening to, um, these really creative ways that this data is meaningful.
46:01It's, it's like a whole new world.
46:03I, I have another customer that told me that he feels like going to Highark was as big as moving
46:11from the drafting board to the computer.
46:14Right.
46:15Um, now that dates, that dates him and me at the same time, because we're the same, but, um, I
46:21feel like that's where we're at, right?
46:22Like we're just now understanding how this data, uh, can, can really, um, flow across our business and start to
46:30connect things.
46:30It's a great example.
46:32Yeah, definitely agree.
46:33And I think one thing we didn't really talk about, but I, I think building something, whether it's a process
46:40or whether it's a spec doc or whatever, like that feels hard, but really it's not.
46:45Building it is easy when it comes in comparison with the maintenance of it, because as soon as you put
46:51something out there and then you need to go make a change.
46:53And then again, my example of 20 houses you sold, you proliferated it and, you know, you have to make
47:00sure that the systems that you're building are not only work when you turn them on day one, but it's
47:05all configurable with your variables and patterns.
47:08Like you say, so that when that change inevitably comes, it's easy to maintain, um, and, and roll that out.
47:15And so where, I guess what made me think of that was, so I'm working on a bunch of, you
47:20know, these skunk work projects.
47:21I call them my son's working on some out there.
47:24Uh, but they work.
47:27The scary thing is they work.
47:28And then this other scary thing is you need to move those into some sort of, you know, production environment
47:34so that they can be maintained.
47:35Cause when you have the next great idea, that thing just needs to be running like clockwork.
47:39So we haven't, we haven't rolled all this stuff out into a, you know, production agent environment yet, but that's
47:46also on our radar.
47:46So you can save that for the next podcast.
47:49Yeah.
47:50Yeah.
47:51We do the same thing.
47:52You know that, right?
47:52We, um, we have a, we have a testing environment and that's really important.
47:57I guess, you know, sometimes failing is as valuable as, uh, as succeeding, right?
48:02Yeah, for sure.
48:04Um, Lauren, I personally have a question for you.
48:07I, I feel like, you know, I, I was giving a presentation at IBS and I, I asked, uh, the
48:13audience who feels intimidated by, um, taking the first step towards, um, you know, some of this type of, uh,
48:23AI functionality in their business.
48:25And really everybody, like the, the room, pretty much everybody raised their hand.
48:29Were you intimidating, intimidated by some of these things?
48:33How did you, um, take the first step towards, um, this, this current state that you're in?
48:41Probably.
48:42I feel like I, uh, I was more excited than intimidated maybe because I come from more of a tech
48:47background, but, uh, and I would also say curious and skeptical because, you know, there's a lot of buzz, a
48:54lot of hype.
48:55Like AI is probably overused word, but I was definitely excited to go look.
48:58Um, I just did what everybody does.
49:01I went and downloaded, I signed up on chat GPT, probably use that for like three years or maybe, I'm
49:06sorry, maybe like a year.
49:08Um, and then I got introduced to Claude and I'm not getting paid by anybody out there, but they are
49:12two different tools for two different purposes.
49:14Um, but I think, and then I just listened to a bunch of podcasts, so it's not, and I'm not
49:19a developer, I'm not a tech guy.
49:21Um, so humans that are not tech can teach themselves and or take classes or watch videos or all that
49:29good stuff.
49:29But I think you just jump in, get your feet wet, start small, don't put it on something that's gonna,
49:36you know, take down the company, but just get started.
49:39Yeah, I love that.
49:40Well, and I would be remiss if I didn't point out that you also, you know, reached out and researched
49:46and contacted Highark as, as part of that, that, uh, adventure that you're on as well.
49:51So, um, love your, your outlook and your attitude overall, Warren, appreciate your willingness to take a chance on, on,
49:58uh, a lot of things.
49:59It sounds like.
50:00You bet.
50:01It's been an awesome journey.
50:02I look forward to what's next.
50:04Thank you for your time.
50:05Great.
50:06Allison, do we have a couple of questions?
50:08Yeah, go ahead.
50:09I have a couple of questions that I'd love to post both of you.
50:12And I want to start with how should builders think about prioritizing AI investments versus other technology initiatives?
50:23Well, I, I'll give a quick response to that.
50:25And Warren, you probably have a better, uh, more meaningful response, but I think that, um, what I would recognize
50:31is that a lot of other technology, um, relies on data that humans have to create manually, uh, without AI,
50:41right?
50:41So I think that you might be doubling up your work if you don't recognize what AI can do for
50:48you.
50:48Um, and you spend a bunch of time creating a bunch of data for a different software, um, you're going
50:56to get, you're going to get to the end of that.
50:58And I think look at AI and realize, oh man, I wish I would have had this at the beginning.
51:02That's just, that's just my opinion.
51:03I don't know, Warren, if you have any other thoughts.
51:06Um, I think like having clean data and well-defined processes is, is mandatory.
51:14But if you wait until you have clean data and well-defined processes, you're never going to get started.
51:19So I think you got to do both in parallel.
51:22And just, I don't think it's, investment in AI is pretty broad term.
51:28I think you can get started for free pretty much if you're using chat or cloud or something like that,
51:33just to test it out and see how it might work in your environment.
51:36But, um, I think for long-term success, you absolutely have to have like two parallel paths going because if
51:42you get something working also in AI, but it's operating off of poor rules or poor data or poor patterns,
51:48then you're not gonna be very happy with the outcome.
51:50But, um, but it's not expensive to get started.
51:54That's actually a fantastic segue to my second question for both of you, which is for builders who are AI
52:02curious, but haven't taken the first step, where should they start?
52:08Yeah.
52:09I mean, I think Warren's given some really good examples.
52:12And, um, you know, I think that a lot of times people are just like afraid to try things, right?
52:19I, I, almost every time I see a new, like not literally every time or else would be my, the
52:24only thing I ever got done.
52:26But a lot of times when I see something in my feed, you know, that's like new or a different
52:30way of looking at it, just try it, right?
52:33Like you're, you're not going to break anything.
52:35And, you know, if it doesn't work the way that you thought it should, then what's, what's really lost?
52:40Do you know the answer to whether or not it's for you?
52:43But like I said, I, I think that, um, the big picture answer is thinking, think about what problems you
52:51need to solve first, right?
52:53Like that was the point of my kind of framing up with this conversation is that, uh, you're going to
52:58waste a bunch of time looking at a bunch of over overwhelming things.
53:03If you don't give yourself, um, a target, an outcome, right?
53:08Like I need to get better about, um, addressing this pattern that's emerging in my business.
53:15Definitely agree with Jennifer on that.
53:16And then I would say, uh, building on that, like once you know what you want to solve, the good
53:22news for somebody who doesn't know anything about it is like every medium that's out there.
53:28It has great free stuff.
53:30If you like YouTube, if you like Instagram, if you just want to surf the web, podcasts, books, like there's
53:37a million different ways that you can get very digestible, very rudimentary information.
53:42So I would say, just pick the one that you like the best, go out there and, you know, do
53:46a search and watch a video and you'll probably wander eventually into what you want.
53:54Now, my last question for both of you is about buy-in, which is a pretty critical part of this
54:00whole conversation.
54:02How do you get team buy-in when you're introducing AI driven workflows?
54:09Warren, I'll let you take that one.
54:10You probably have a much better perspective than I do.
54:15Are you really enjoying the way that's going for you out there?
54:18How about we try something else?
54:20So I think, uh, I do.
54:23I mean, I, I think definitely when we first introduced Highark, we were two years in the past and we
54:31were probably a slightly different company.
54:32And there was a lot more resistance to this new fan bangled technology.
54:37And we're home builders.
54:38We're not, you know, we're not techie.
54:39And I would say two years later, like after they see it work and they, you know, put the effort
54:45in to actually get their hands on to build it.
54:47But now they're a lot more open to, you know, just technology ideas, whether it's like Slack communication channels, just
54:55like things like that.
54:57But I feel like, I think that people understand now that unless we want to hire 50 people to grow
55:03another $25 million in revenue, which we don't want to do, um, we have a really nice profit sharing pool.
55:10So everybody's like, okay, great.
55:11If we can do more with less and love our jobs and be efficient and be proud of the work
55:16and the company, like that's our goal.
55:18So I feel like with that, for us introducing these types of things is, is getting easier and easier every
55:24day.
55:26That's really inspiring to be honest, Warren.
55:28And it, it makes me think about like a lot of times when I'm talking with builders, um, and individual
55:34users within the builder, I say to them, you know, you realize that this automation is useless without you, right?
55:41Unless you can identify something that you want to offer to the organization, or you want to take off of
55:48your plate, then it's just, it's just a fancy tool, right?
55:54And so I think that's what, um, I, I try to push people to understand is that, you know, and
56:00you can usually see it on their face when it's like, oh, I don't ever have to draw a corner
56:05board again.
56:06I can just like magically make corner board appear every time it's supposed to be there, right?
56:12Like they're like, yes, that's, I'm, I'm all for that.
56:15Right.
56:15And so I think finding that, um, not just for your organization, the pattern that, that exists for your organization.
56:22I think it's also challenging your people to find the patterns and the things that they do every day that
56:28they just don't, I, you know what?
56:31I don't need to keep drawing corner board.
56:33I can kind of outline when it's required and then my, my tool can do it for me as a,
56:39it's a good way to look at it.
56:40Yeah.
56:41Yep.
56:43Awesome.
56:43Well, I know we're about at time.
56:45Um, I just want to personally thank you, Warren, for taking, uh, uh, time out of your day to chat
56:51with us.
56:52Um, certainly have appreciated your perspective and, um, really, um, a big fan of, of what you're doing at Ashlar.
56:59And I know that you guys have come a long way and, and you have a great team of people.
57:03So, um, really nice, nicely done.
57:07And, uh, we, we can't wait.
57:08Let's do this again next year.
57:10And we'll hear more about some of the, uh, the things that you're just getting started on.
57:14Absolutely.
57:15Appreciate you.
57:15Appreciate higher work, higher and all the support.
57:17And, uh, it's been a lot of fun looking forward to what's next.
57:23Jennifer, Warren, thank you so much for joining us for today's discussion.
57:29I feel like I have a couple pages of notes.
57:31I'm excited to see this conversation in the future.
57:34AI moves so quickly.
57:36Jennifer, you said that.
57:37And to our attendees, thank you for joining us today.
57:40There was a question about whether or not you will get a recording.
57:43Yes.
57:44I promise you will all get a recording of today's webinar.
57:47You will be able to go back and revisit your favorite parts and send it to whoever you would like.
57:52Thank you again for joining us.
57:54And we hope to see you all again soon.
Comments