Why Your Firm’s AI Adoption is Failing (And How to Fix It)

I I we see this [music] every day. It's stalls between okay, here's the proof of concept. This is great. This is what it can do. Between there and implementation [music] where you have, you know, a good portion of any firm or group or company using the technology. >> [music] >> What gets in the way in that that gap that we see? >> The fact that >> [music] >> systems are, you know, siloed or they've got disparate or fragmented data is something that is usually a big showstopper [music] inside of this. You get to a spot where we just don't have access to the data or we just can't find the data or where the system hasn't been indexed or not interoperable. >> [music] >> The change management aspect of getting this adopted afterwards, that part is the tough part. >> Many firms in the AEC industry are experimenting with AI. Right? Pilots, testing tools, and exploring what's possible. But somewhere between proofs of concept and implementation, momentum stalls. Technology is there, interest is there. So, what is getting in the way? Today's guest has a front row seat to that question. Benjamin Massa is the chief technology officer at Newforma, a SaaS platform serving over 500,000 users across the global AECO industry. With 15 years of engineering and technology leadership spanning startups, mid-market, and enterprise organizations, Benjamin brings a rare combination of technical depth and business acumen to every decision he makes. In this episode, we're getting into what it takes to move AI from concept to real-world impact, how to scale engineering teams without sacrificing culture or delivery quality, and what the future of driven decision-making looks like when project information is no longer siloed. If you're leading a firm, managing projects, or thinking about where technology fits into your growth strategy, this episode is for you. But before we jump in, I want to tell you about AECPM Connect, a series of in-person events we created at EMI for AEC project managers and leaders who develop them. You want to stay on top of PM trends, elevate your team, and deliver stronger project results? These full-day events are built for you. We're coming to Kansas City this October. Learn more and register at aecpmconnect.com. I'll be there and I hope to see you there as well. Now it's time for our conversation of the week with Ben Bassoe. Ben, thank you so much for taking the time to join the show today. Welcome. >> Pleasure to be here. >> So, I think we have a we have a nice interview set up, Ben. I know we were just chatting a little bit earlier about your unique perspective um coming from a different viewpoint here in the industry. So, we'll just get things kicked right off. So, Ben, please tell us a little bit about yourself, what you're focused on in your role as CTO at Newforma. >> Okay, well, I'll start by getting a little bit about who I am in a nutshell. Um, you know, for me, very much so, I come as a builder. I've been in the software engineering space for about 26 years. I've had a couple of my own startups. I've done a whole bunch of different spaces in in the past, everything from adtech to marine to military. And I never really touched too much into the construction technology space, but I've been really enjoying that space coming on board as the CTO because it's been a space where I've been able to take a lot of the tech that we've, you know, that I've used in other industries before and been able to see opportunities for to be able to apply that inside of the construction tech space. And so, loosely, my role as coming on board of CTO, which I've only been at the organization now for about 10 months, has been taking a lot of what I've learned over the last plus decade or so in the AI space and taking that AI knowledge and being able to apply that to some of the workflows in the AEC space to be able to optimize that, make that a little bit better. So, essentially, in a nutshell, taking AI, translating that into the construction technology space, and how can we use that more effectively? And not just building vaporware or building AI for the sake of calling it AI, but actually building stuff that is really useful for real people, cuz that's where I get a lot of delight in uh, bringing smile to people's faces. >> Excellent. So, I know, you know, Newforma's, guys are pretty sizable company. You have pretty sizable user base, right? Over 500,000 users across our industry, AEC EO. So, when you're making decisions at that scale, so how do you guys figure out what matters to the end user? Um, because that's really what this is all about, right? >> That's, yeah, for sure. So, I think one of the things that's been very good thus far as an organization that I've seen, uh, that we do particularly well is we listen really well to our customers. And so, even though we've got hundreds of thousands of folks using our system, uh, we we meet regularly, uh, doing what we call a PAB or a PAB. So, it's a product advisory board. And we do that at multiple levels in multiple locations around the world to be able to kind of listen to our our customers about what's working well, what's not working well. Getting that as a as a as a, you know, point of input to be able to understand what we're doing, um, you know, so it's kind of a sample set. Now, these are champions and they usually have a lot of influence on on in the industry. But then, I'll flip that over. There's other stuff that we do as well, too. So, we're also a data-driven organization. We keep metrics around how people are using our applications and we use that to be able to tailor, um, you know, the the improvement over time. So, we see if somebody's not or in users are not using a particular feature inside of our application or they're spending too long on particular operations, we're constantly tinkering with that to be able to optimize that to make them better. We do do a lot of interviews as well, too, that are kind of with individuals just to kind of see and and, you know, kind of dig down. But then, I'll flip that around, too, on the scale side of thing. This is something that speaks near and dear to my heart, uh, where performance is really an important um, um metric as well, too. So, this is about actually us inside of the engineering teams building things that are highly performant. And so, even though when we're building each individual feature, we are iterating and starting with a small little sample set, we're also thinking big and being able to build out the test suites and the scalability around that where we're testing and then hitting it with hundreds of thousands of users to see, does it actually meet our standards for, you know, what would an acceptable performance in terms of wait time would be or being able to perform certain operations. >> Excellent, which is great because, I mean, you guys, like anyone customer has their own vision or opinion, but you guys literally see the activity of your entire customer base, right? So, Ben, you also own the AI product roadmap at Newforma. So, I know, right, that includes agentic workflows, smart search, and any of those other AI-oriented capabilities that we've we've already discussed. So, how do you translate like the idea behind AI into something that people can actually use? >> Yeah, I mean, I think smart search is a great starting point, right? And the thing that I always say um to to folks is that, you know, we we've, you know, search has become commoditized in many fronts where, you know, before it was Google and you nowadays a lot of people are using LLMs to just search what they're looking for. And they've done so very much and what's changed dramatically is that allowed people ask questions and they search for things not in a keyword search, but they look for things in a semantic search. So, like, give me all the issues that have occurred in the last 10 days at schools, you know, in some in their system. And they're like, okay, that it's smart enough to be able to pull out that information. Now, in order to do that from a smart search point of view, you can use AI uh to be able to, you know, retrieve that information more intelligently these days. So, even though search has been, you know, you know, people thought that this was a done deal and this is something that was, you know, we we you could actually um it was it was already a solved problem, it's continuously getting better. Uh and when we come to this the semantic search side of things, there's still a lot of improvement to be had there. Now, why do I start with the smart search side of things and why do we call it smart search? It's because it's built on top of an agentic framework that allows that that is actually indexed differently inside of the system so that it permits for you to do those semantic searches. It's not just a keyword search, it's not just by category, it's allowing you to do things where I'm speaking to it in almost natural language, and giving it a question and it's now being able to retrieve that. Now, you're like, "Okay, well, big deal, you know, this is a solved problem, this is something that our LLMs do already every day." But, it's a foundational element to be able to do smart things inside of the AEC space. Now, why do I say that? It's is because before you can do anything intelligent with the information, you need to be able to find the information. So, imagine and obviously right now, we do we have demographics around how long architects and engineers spend looking for information inside of their system, and it's it's actually it's really surprising to see how much time they waste actually just looking to be able to retrieve the information that they need to, you know, grapple together or combine together from multiple systems that are siloed and be able to bring that into one spot. But, now the fun part starts. Once you've actually found that information, then you can start building on top of it. You can start building now, which has become very popularized in the industry around agents, right? You can now start building smarter agents that do things with the data that you've now retrieved and automate a good portion of those workflows. That's a huge that's a great starting point, but that's definitely something from the, you know, AI roadmap that we're actively working on right now and continuing to build and augment on, but another good reason from an engineering point of view as to why this is a smart approach is because now it gives you a lot of extensibility. So, it's a good engineering practice to build something that then you can extend later, and this is where you come for that agentic framework. Once you're able to find information, now you're able to build different types of agents on top of that to be able to go and build. So, an example of one in our space, we have a submittal agent that we're working on right now. And what that does is really it automates a good portion of the information from when you're actually doing a submittal. We have certain specification books inside of our system that you must adhere to for certain materials or certain building codes that need to be respected well when you're doing that submission. What we're able to do with that is automate that away. Not only just retrieve the information that's pertinent to the user, but automate the fact that we're cross-checking the two. Does this actually give us a thumbs up and it meets the requirements or doesn't it? And what that is end up what that ends up doing is that we're not replacing any of the users that are there, we're just augmenting it so that they can pass their judgment more quickly and save them a significant amount of time. So, these are the just an example of the types of agents that you can start to build on top of the information once you're retrieving it in this new way that I just mentioned. >> And this story, of course, goes down the down the line to adoption, which is one of my one of my favorite topic. So, we I we see this every day. It stalls between okay, here's the proof of concept. This is great. This is what it can do. I we see this every day. It stalls between okay, here's the proof of concept. This is great. This is what it can do. Between there and implementation where you have, you know, a good portion of any firm or group or company using the technology. What gets in the way in that that gap that we see? >> Yeah, okay. And this one's a very I mean, I could probably talk for hours on this one, Um, I'll I'll keep it short. There's a bunch of ones that really, you know, get in the way really quickly and I'll leave those out because they're a little bit less interesting to talk about or a little bit more mainstream. You know, the fact that systems are, you know, siloed or they've got disparate or fragmented data is something that is usually a big showstopper inside of this, meaning that they actually you get to a spot where we just don't have access to the data or we just can't find the data or where the system hasn't been indexed or they're they're not interoperable. So, those are the kind of things where I'll leave that one aside because technology in the question that you just mentioned is usually the easy part. The change management aspect of getting this adopted afterwards, that part is the tough part. And so, what I see often stalls inside of these and I'll tell just like a little small little quick story to give an example of this. Often times what happens is that you get this smaller team inside of the organization that are like, "Wow, there's this really cool idea that we should be doing. Let's prototype it." They ended up prototyping and then using an LLM, they built it fairly quickly and they're like, "Wow, this thing is the best thing since sliced bread. This is really great. This is amazing. Everybody should be using this. It's going to save everybody time." They never went and got a a champion. They never got a a buy-in from the rest of the organization. And so, then they start taking this best idea that they think ever and they start pushing this into the organization and they're like, "We don't need this." Or "We're not ready for this now." Or "You never talked to us about this before." Or And so, you get the whole change management 101 inside of an organization that people start to push back on this, right? And this Oh, that's a new way of working. Why do I need to work this new way? I've always done it this other way. And so, this is those are the kind of things that really push back hard on that. And so, despite the fact that this was really a genius idea, it's a great thing for the organization, they get pushback. And so, what I strongly suggest often to getting over those hurdles is really about getting that awareness of what's going on as early as possible, getting some champions, and there's a lot of change management 101, you know, I would say tactics to be able to still get your AI through the door and being getting it out there and so you being used. Uh and often times, um you know, that stems from the fact that, you know, just getting more people on board and actually seeing the benefit of it is a great start. >> Transform your technical experts into impactful leaders with EMI's Engineering Leadership Accelerator. [music] Tailored for AEC professionals, this program enhances communication, delegation, and decision-making [music] skills. Flexible, interactive, and designed to fit your team's schedule. Elevate your team's leadership capabilities today. Visit engineeringmanagementinstitute.org and click on corporate training. >> Excellent. And I really appreciate that answer because, I mean, we've all have that experience, right? You say you just go build a very simple agent, people love it, but there's so much more that goes into getting that deployed successfully to an entire department, an entire company. And and really it's often times the leadership that needs to be on board who may be like the least educated about what's being built and the purpose that it serves. So you have to get buy-in from a couple of different parties. >> Yeah, I definitely agree with that. I mean, I think one of the other thing, too, is that when your leadership is already pushing back against it, it's already going to be something that is very, very difficult to to to overcome inside of an organization. But I I definitely think that what's really helping though these days, at least at the leadership side, is that a lot of organizations have grown this anxiety about AI taking over the world. If we don't do it with AI right now, you're going to be left behind. So there is definitely is much more awareness and openness to be trying these things. Um but at the same time, there still is a little bit of change management 101 that gets applied to whether or not Okay, you know, these pilots succeed or fail. >> Absolutely. And uh again, it's often times the leaders who they may hear all of this talk about artificial intelligence, but less often are they hands-on and trying it themselves. Outside of, right, like engineering and innovation technology space. So, it's always great when they can gain that first-hand experience for themselves, cuz that it actually better guides their decisions than something generic they may have read in online or heard from a peer. >> Yeah, the other thing too, I think that really helps in this is that, you know, getting um getting a culture that adopts the idea from the get-go that it's okay to start pilots, it's okay to start little proofs of concept. Um but the one thing that I've noticed um through my experience is that a lot of these people find it easy to build a proof of concept, but then coming back to your earlier question around scale and enterprise grade, they realize that, okay, it's easy to get this proof of concept off the ground in a day or two. I've I put something, but then scaling it to a larger organization, creating enterprise grade software, or sometimes just as simple, and I know this will come across as funny, as soon as they hit their first bug, they're like, uh I don't really know what to do anymore. And they start Yeah. >> This is a topic I love discussing because I think so you, right, you know, vibe coding, the idea behind anybody can build software. So, for that first 10 to 20% that proof of concept is awesome. But as soon as like you said, anything in life, you hit that first wall, it's like, uh-oh, what do I do? It's and I have done this myself multiple times. You spin up a quick proof of concept, it takes a couple hours, but there's a reason that there's a reason we still have professional software developers. It's It's In my opinion, it's different when you have those, let's say, by going tools, cloud code, whatever you want to use, in the hands of a professional. That's different. But, for the layperson, right? Like, there's a reason we're not professional developers. It's not that you shouldn't try, just understand the difference between a pilot and a proof of concept and an enterprise piece of software that has to work for hundreds, thousands, tens of thousands of users. >> Yeah, definitely. And I I I think that there's, you know, you know, I had this debate the other day. I was at a CTO, uh, you know, kind of gathering, and we're around the table, we asked these questions, right? Are very much around the the notion of when do you think the modern the the way that the full-stack software developer these days, when do you think that that job will go away? And we had some people answer across the table, and like the the the the first one was 5 years, 2 years, 2 years, 5 years, never. Right? And I was like, okay. Well, so so then we started exploring that and talking about that a little bit more. What do you mean by that, right? And it was kind of fascinating to see about the different perspectives that people were sharing at the table. But, to to your point, there are certain portions of that job that are definitely going away these days, especially when it comes to translating, let's say, business logic into code, just purely. This part is definitely going away faster and faster. The coding agents are getting better and better. And us internally at New Form are we're using this right now. We're using factories to be able, you know, coding factories to be able to code our our features in our applications. And it it is definitely augmenting the speed at which the the the team is able to go. But, then there's the other aspect of sides of things where it comes to scale, and it comes to operations, and devops, and the bugs, and things that get introduced, you're still going to need the judgment of what a good software engineer can put in place as good practices. And so this is very much and you've heard a lot of folks talk about this these days is investing inside of your AI harness or investing inside of your AI infrastructure, right? Around that. And that's where I think if my advice on that is very much about giving organizations visibility of what you're building with your AI is very key. So for us in the example of this, you know, leveraging on top of technologies like something like LangGraph is it gives you the ability to build agents, but it gives you also the comprehension to understand how they communicate with one another. So that when things go wrong, as a human, you can now jump in human in the loop and debug it to be able to understand, okay, this actually is failing because of this. Let's optimize that. Let's create a skill for this or let's be able to modify this so that it it knows next time not to do the same error or it knows next time that it it shouldn't do something. And so that and that's how you grow your system to be able to improve and get better over time. But this is definitely something that I see come back as a pattern again and again. I love the the notion of being able to iterate quickly with those proof of concepts. That's definitely helped a lot, especially on the product discovery side of things. But at the same time, bringing that into enterprise grade that works for, you know, in our case, 500,000 users, um those are the kind of things where it's not as easy as a problem sometimes to solve and you definitely still need the skill set that has been there for, you know, that you master over years. >> Yes, I don't Yeah, it probably that says uh please work, no bugs, 500,000 users. Yeah. May maybe not as simple as it seems. >> Exactly. I agree. >> And so, you know, going going back to data-driven decision-making. So we talked about how you guys use it internally to prioritize feature sets, maybe simplify some features. But for the user side, we often find come data silos, right? Across different stakeholders at a project. So, I know A, somebody else knows B, but I don't know that they don't know what I know, and I don't know what they know. Cuz I don't know. Um that little that little word soup there, but essentially how does what you guys what you guys are doing change that? >> The the thing is is that one of the the patterns that I've been trying I mean that we as adopted as our I don't know as almost as our DNA inside of the organization is that we're very we very much adopt a conduit pattern. Okay, and so the conduit pattern is something where hey, you're just the conduit layer between multiple systems that and so you the idea is you don't try to consolidate it all in one system. You don't try to go and suck in everything and and be the one-stop shop kind of what used to be the the trend in the past around creating one data lake or one source of truth. You now have multiple sources of truth and you're a conduit and you make it easier to be able to go and probe and get the information from the information where they lie. That makes it more extensible. It allows you to augment our time and what you and it allows you to if you don't know something out there it the conduit pattern also allows you to use something that's more you know, I'll use the technical term, it's more of an observer pattern and the observer pattern is kind of you can broadcast something out there and say, do you have this information for me? Do you are and so you can kind of build the sense of awareness out there inside of this pattern to be able to to be used. Now, if you permit me, I'll go a little bit really like where the future is going with some of this stuff in terms of the the the agent agentic frameworks that are going to be used in the future. There's something that's becoming more and more popularized these days around something called agentic farming. And agentic farming is like creating multiple agents that do one specific task, but with them they've got a sense of discovery between them. So you basically broadcast like, "Hey, I'm an agent. I can do this. I'm able to have these capabilities. Here's how I can communicate." and etc. and you're giving a list of my what can I do? And there and then if the and then you've got hundreds, sometimes even thousands of these inside of a farm. And at the beginning of this farm you're saying, "Complete for me this task." And those those agents figure out between themselves what should I need? What do I need to be able to accomplish those tasks? And at the output it's spitting out the tasks that it's completed. And you're like, "Wow, okay. How does that work?" Well, you need to have that level of discoverability built into your framework when it comes to your agentic farming and how they can achieve something. And that's where the trend is going because what's also happening in there is that if ever you don't have a capability inside of that agentic farm, the farm builds itself. So like, "I don't have an agent that does this. Build me another agent that will actually do this." And so as you can imagine it's augmenting itself over time. Now, there's a whole bunch of security concerns and a whole bunch of a well, is this going to go wild wild west? Is this going to explode? And of course it will. But if you've built the proper sandboxes and enclaves around your data, around what you can do with it, and the governance on the security side of things, and observab- and the observability and monitoring, then it's definitely the trend of where things are going, um you know, over time. And I know that probably didn't exactly answer your question, but it kind of gives you an idea of um you know, of where my head is at around some of that. You know, the the whole notion of not knowing what other people know. There's another pattern as well, too, that's used quite often out there and it's through the notion of clean rooms. And so clean rooms as a service is kind of this thing where a lot of industries do this where they're like, "Well, listen. You're kind of my competitor or you're my you know, you're somebody that we work in the same industry. I don't want to share my secret sauce with you. I don't want to share, but I I to share some of my data so you I can get insights and usually they swap. So it's kind of like, "Hey, you're architecture firm A and I'm architecture firm B. I want to share what I'm doing to learn from you, so I'll swap you some of my data, but what the clean rooms allow you to do is anonymize that data." So it's the kind of thing where it's not revealing necessarily any of your secret sauce, it's giving some of your data there, it's anonymized data, but it's stuff for them to learn from and vice versa. So usually they're doing that in the form of exchange and sometimes they're doing that in terms of a a clean room where they're actually selling their data to be able to get insights. But on that notion, the AEC space is still very, I would say, new to the concept of trying to think about how do they monetize the data that they're sitting on top of cuz they're often times sitting on you know, 20, 30 years worth of data that's just sitting there and it's not really doing anything for them. Um and so they haven't thought about how to monetize that either internally or externally yet. Now, most of them know that it's there and they know that they can do something with it, but they usually have not quite gotten over that hurdle yet to be able to get things in motion. >> Yeah, and I think when it comes to the AEC space, you know, any mature industry, in my experience, we're pretty tight to the chest with with our data and sharing, you know, best practices beyond what's already published industry guides isn't really super common. But to your point about not knowing what to do with it, I mean, like if you don't have somebody who's one thinking about this on your staff, which a lot of leadership teams are, but then how do I execute and what does that look like? But what's really cool and when I, you know, talk to people about similar topics is like you can go work with off-the-shelf tools and with the right, you know, with the right setup, permissionings, and adding the data that you're working on in an accessible spot, you can build yourself a little proof of concept. And if you want to go hire a professional to take it to the next level for you, well, better than starting from scratch. >> Yeah, I think one of the hurdles I've noticed is that a lot of organizations don't take the time to invest into, you know, a role such as like a data scientist. And I I say data scientist, but that role's very, you know, I would say it varies more enormously across organizations. There's certain data scientists that are very much into the machine learning side of things and they're doing but at the end of the day, somebody who's taking a look at your data and actually thinking about how can we start to derive insights from it and doing that in an intelligent way. I think due to what you you mentioned as a point, there's great tools out there to be able to start to do that at a very surface level and get going. Just that investment to start to get going is something I believe is fairly easy to do and but a lot of companies still just have not quite taken, you know, bit the bullet and and invested into that yet. And the ones that have, every single one that I've talked to that have, have hired that for that role, have automatically within, I want to say six months of the within the first year, have rehired another one of that roles just because they've seen the value that it started to bring for the organization almost within a few months of return on investment. >> Yeah, or if, you know, if you, right? Like like let's just say I give an example, right? Like say you have 20, 30 years of pricing data. You're contractor and it's sitting all over your SharePoint, your Dropbox, wherever. So, the way that I would I would tackle it is like if you're if you've got it off the shelf LLM, you can do some you can do some search, get data or or a segment of it in one place, start asking questions about it. What is it that you're trying to figure out? What's the problem you're trying to solve? You may be surprised at how far you get. Then, if you realize that you're making progress, but maybe there's something more you want to get done, you're not exactly sure how to do it. Hire somebody fractionally. You don't have to start out with a full-time data scientist. You can go out, get some consulting done, help them figure out your problem, and take it at the pace that works for you and your organization. >> Yeah, the other thing I've seen work fairly well as well, too, there's some good off-the-shelf tools that, you know, you can go the LLM approach, but if you've got large amounts of data, your context windows might end up filling up very quickly, and so it might have a difficult time, like, you know, you're going to burn through a lot of tokens, you're going to burn through, and it won't be as efficient. So, there's some tools that have been optimized for this, and the the tools themselves allow you to just get early-stage insights. So, what I've kind of found that's a cool hack around that is being able to have external parties come in and do a little mini proof of concept. Show and usually they're doing them for free or for a very minimal cost. Most of them are for free cuz they want to showcase the value of what they can bring to the table, and often times what I've seen is if they've able to showcase the proof of value inside of their they're showcasing of their product, usually that's enough to get it off the ground, and the product itself is usually cheaper than going off and hiring somebody, but it's a good starting point. And then once you get there, then, you know, you start to think about, okay, do I staff for, you know, a project or a program around this? >> Oh, and that absolutely. And if and I like I like the LLM approach because if you already have one and it's general enough to get you started, well, after it takes a look at that sample set of data, you can start asking it questions about how do I take it to the next step? Maybe it identifies one of those providers. Maybe it finds an expert like you that you end up reaching out to to just ask a couple of questions. So, it's cool cuz it it's it's kind of the analogy to bike coding, right? Like, get that proof of concept, get it off the ground. When it comes time to bring in the expert, well, you've already got a couple steps in the right direction. >> Exactly. And often times you discover stuff there that you didn't even know existed. And there's like see like I've done broad things of like, "Can you give me insights about what you're what you learned about this data?" Right? And just at the LLM level with, you know, let's just say 1,000 plus documents, not a huge amount, but they already came back with patterns and you're like, "Damn, I never even saw that." You know, like and so it can see stuff that you just don't process quite the same way. >> Well, and what I love about keeping it in the SME's hands and keeping the tooling set simple is because like as the SME writer business owner, you're going to be able to sniff out BS better than anyone. >> Oh, yeah. >> your business. You know if an insight comes to you, right? You could pretty quickly figure out if it's plausible before you just start going down rabbit holes that aren't realistic. So, I love that because that it equips you as the non-technical SME to get in, start doing some work, and then better educate yourself before you go say make a more expensive purchasing decision. >> Makes sense. >> So. >> It's and it's all right, there's never been a better time, but there's also there's always a place in time for the Xbox, so just just remember that. Ben, this has been this has been awesome. Last question for you. So, if you're a firm looking to adopt AI or any new technology, what do you tell leadership teams to focus on to make the transition from hey, like this is where we are today to whatever vision they see enabled with that new technology? >> Yeah, okay. So, this one I would say this is something that we've done at Newforma and it's also something that is near and dear to my heart is that you really need whenever you're doing these types of transformations and starting to adopt something, especially at the pace at which AI is moving today, and especially at, you know, I'd say that you need to give your teams space to be able to take the time to learn. You need to be able to give them some space to be able to So, what we do is we give everybody on our engineering teams, a you know, half a day to a day a week just to be able to go and experiment. To be able to go off there and learn. And so, yes, we've put in place training programs, but giving them the ability to have time to actually apply that to real use cases. And I call that play. But the idea of being able to play with the tech to be able to tape what you may have to read or learned and apply it to a real use case is something that is is key. And it's not just about hey, we're giving people some free time off. There, you know, to to just go and learn on their own. We have it set up in a way where they're sharing that information amongst themselves as well, too. And that's almost mandated. So, we want to be, you know, share casing showcasing a little bit of what we've actually learned so other people can see some of the cool stuff that's going on out there, learn from it. What I've learned from this as a piece of advice is that the champions were the smart geniuses that are going and leading the way on things are not necessarily your best teachers. So, this is the first thing that like they actually want to just go faster and they like somebody who's a little slower trying to get adoption, it's not necessarily hey, team them up with your smartest person. There's so there's certain people that are great at being able to teach, so identify those and make sure that they're kind of giving back to the team. And the other thing from the leadership side of things is lead a little bit by example and spend a little bit of your the portion of your time as a leader actually teaching. And that is actually key as well, too, to getting some of this adoption because it shows not only that you care, but it also shows that you've got some of that thought leadership. It's moving so quickly, you're never going to be able to be a leader in all spaces, but what I do especially in my space, even though I've been branded as an AI expert, I'd say that these these are the kind of things where I know that that doesn't really exist anymore. It's moving so quickly, but you want to be able to learn certain things to be able to teach that back to the team and be able to provide some guidance on it. So, take the time to invest into being able to teach the team. And in order to teach, that means you need to learn it yourself. And so, there are certain things there where that's really key and important inside of the from a leadership standpoint, that you're not just talking the talk, but you're actually walking the walk as well, too. >> Absolutely. And right with as broad as this space is, there's just right just like human expertise in any area, right? There's we've only we only have so many hours in the day. There's only really so much of that narrow band that you can focus on and become an expert. Just kind of is impossible, unless you're like one of those cool polymath people who could just do it all, right? But I know I'm not, so. But it's really been a pleasure to get to to speak with you, get to know you better, and I'm sure the audience got a lot out of this episode. So, if they'd like to reach out to you, ask you additional questions, or just chat, what's the best way for them to reach you? >> I mean, I'm somebody who's pretty open and transparent. Reach out to me on LinkedIn or my my email. Happy to share that, you know, on side of on this side as well, too. You know, so and you know, those are the kind of things where I love to talk shop, as well. So, if there's folks that want to reach out and, you know, dive a little deeper in some of what we've talked about today, happy to do so. It's something that, you know, it gives me a lot of joy. I I organize, you know, some of the CTO meetups in my area and so forth. I love to give back to the community and kind of give back as a as a whole, just as a technologist in general. So, happy to share with my email as well, too. >> And thank you so much for taking the time to join us. >> Thank you very much for having me. >> Absolutely. Take care. Bye-bye. Please remember, you can find the show notes for this episode at aectechpodcast.com. There, you'll find a summary of the key points discussed in today's episode, as well as links to any of the resources, websites, or books mentioned during this episode. Until next time, I wish you the best in all your engineering and technology endeavors. >> Mhm.