The Construction Leader's Journey: Updated Insights & Industry Trends

(1:30) The survey evidence and the findings that we’re gonna share with you all today. This is how you can use that data to translate it into five practical workflow examples and the leadership you need to make them repeatable. And here’s the core question that we all strive to answer where and how can this data be used to better change your project performance and improve your cash flow. Before we get into the details, here’s some quick housekeeping. Yes, this is being recorded. It’ll be shared out with everybody afterwards, including the report findings and the content. Also, we have obviously two speakers. We have a dedicated Q&A session at the very end. We’re gonna park about… 15, 20 minutes for that at the very least. But if you have a question that you’re compelled to ask, please just ask it. Two ways you can do that, you can raise your hand using the Zoom functionality or there’s a Q&A chat box. Forgive me if I’m using the wrong terminology, but there is a Q&A function. Just submit your question if we can address it ASAP and organically as we go along. We absolutely will.

(2:41) Let me share with you and flash up our agenda. Of course, we’re gonna get into introductions. We’re gonna get through the content, updated insights, as I mentioned earlier, five use cases, and of course your path to AI maturity, and of course the Q&A. And as a quick plug-in teaser for our future installment of this series, we’re gonna dive deeper into the topic of risk, insurance and legal landscape. We’re targeting probably Q1 next year. Paul, Eric, does that sound about right? That’s correct. Yeah, somewhere in the second or third week of January. Yep. And we’re looking forward to that. And the reason we’re really excited about our findings today is this is a growing body of knowledge. This is not stagnant. It’s not going to be stale. This is going to continue to develop as we go. So let me get into some quick introductions. My name is Randall Mauricio. I am TAG CXO’s chief Operations Officer. My job, my goal here is to really strengthen operational backbone, make sure the TAG CXO’s network of enterprise caliber executives can deliver consistently and at scale all the time. Of course, we have two speakers, Paul Theisen and Eric Sanderson who are very well decorated and accomplished. I’ll start first with a quick bio on Paul and I’ll make sure I read this verbatim because it’s worth getting into their history. So Paul’s the CEO, founder of TAG CXO. He founded the company to bring enterprise level executive leadership within reach of mid-market companies, of course, connecting them with seasoned fractional executives. Their jobs are simple. It’s to drive immediate value and meaningful impact. And I see this day in, day out as in my work with Paul, this man really champions the mission of TAG CXO to deliver the clarity, the proven expertise to organizations just like yours to navigate change, to help you navigate change, including the one we’re talking about today, and that is AI adoption.

(4:38) Eric Sanderson, president and founder of Red Rocks Advisors, that’s a management consulting firm serving contractors, engineers, and of course owners across the construction industry. Of course, this firm has deep roots in the trades and construction management consulting. And I need to mention that Eric’s also a PI certified master-level facilitator, and he’s known for helping construction organizations make better, more informed, more confident decisions. Now that’s our bios. However, Eric, I want to thank you so much because this has taken a lot of time, energy and effort for you to collaborate with TAG CXO. And I really want to toss it over to you because you have a history with Paul and maybe just a quick commentary on that history and why this body of knowledge is so meaningful for you.

(5:24) Thank you. Thanks for having me. It’s a unique privilege to be able to do this. And yeah, the Paul and I’s relationship goes back quite a ways and it’s always circled around items related to technology and its application and use in the industry. And as we were talking recently, I’d say it’s recent, but now it’s been a little while about the emergence of AI and how it’s influencing our industry. And we just decided it would be a good initiative for us to collaborate our individual talents and relationships to turn and learn what we could, but then ultimately share that information with others. And we know it’s, you know, there’s a lot going on and it’s moving the industry and we wanted to be able to be a conduit for, you know, objective information. And so, yeah, great relationship over the years and we’re continuing to drive it forward and hopefully add some value to the industry. Appreciate you being here. Appreciate you being with us. And so Paul, maybe we’ll take it away and dive deeper into the updated insights. Toss it over to you.

(6:26) Yeah. Thanks Eric and Randall. I appreciate that very much. You know, this AI readiness research concept that we came up with really was born out of what we saw as a vacuum where we could help fill and that is, solid truthful information about what’s going on, what are the genuine trends, and let’s give it back to the industry operators. And let’s take the time, energy and effort to collect information that they need to have to make better decisions, but otherwise aren’t probably gonna go get themselves. So without a sales pitch, without something attached to this, we’re simply trying to find out what’s going on in the industry and share it with you and be honest about what the trends are and the things that you should be paying attention to. We went back through and when we say updated, we never closed the survey, Eric, and we kept getting responses on the survey. But the message is not fundamentally changing. And there is a need, you know, three years, two and a half, three years into all the generative AI craze.

(7:33) We’re getting better at using it as a population. But there is a very important inflection point right now where we need to move away from personal productivity into job, team, and company workflow productivity. And that is becoming very much more well-defined. What you’re seeing here is we took two slides from the previous report that we did, the AI Readiness Report. And we gave you the top three issues that people are thinking about. On the left side is here’s what we’re doing today, the top three things we’re doing today. And you can see those are pretty much personal productivity items. And in order to really get more value out of this AI, we’ve got to move it towards the business. We’ve got to move it to the work. And that is contracts, estimating, deep research and analysis with all the documentation we have in construction. So what we noticed when we started to study this is that these workflows, contracts estimating, let’s just talk about those for a minute, they are labor intensive.

(8:46) There’s a lot of hours and a lot of people involved in analyzing data, synthesizing data, reading documents, double checking things, cross checking. And why is it labor intensive? Why is that so important? Because it’s high risk. You can’t get it wrong. You’re literally in error correction mode trying to make an accurate bid based on real information and honest experience. And that takes a lot of time, energy and effort from people. That is the perfect scenario for an AI support, right? Where you can do a lot of research and a lot of analysis and reduce the labor of analysis and research and synthesis, right? And now translate your actual human labor more to judgment and decision-making, informed judgment, informed decision-making. So that’s the intensiveness of where we’re heading. And so we don’t have a four box, but if we did, it would be high risk, high labor, and you’d be up in the upper quadrant, focus on those workflows. And Eric will talk about that a little bit later. Okay, let’s hit the next slide here.

(10:04) We wanted to sort of net this out for you and create a financial example if we just automate one workflow. And we’re going to talk about estimating here in a minute. 62% of the survey respondents said they wanted, they believed there was value to be harvested in estimating. And or in whatever their version of estimating is in their culture and their reality. So let’s flip to the next slide. So, so we did a what if scenario slide if we just took this as an example. If you’re current in this, these are just making assumptions, you can plug in your own assumptions for your organization here, but let’s just say that We have a win rate of 20%. One in five bids we win. That means 80% of the bids that you’re working on that don’t turn into revenue, the one bid that wins has to pay for that plus make a margin on the business of profit.

(11:00) And if you could utilize AI in such a way as to choose better bids, inform those bids better, get down to the truth, understand your costs with more accuracy, understand your opportunities with more accuracy. Could you then, if given a fair field of you using AI better than your competitors are using AI, can we turn that into two out of five? That’s a good goal. Just pick your number, pick your win rate that you want to improve to, set a target. We want, in our example, when we go from one, we’re gonna get to two. That’s double the win rate, double the revenue, probably more than double the margin if we’re gaining efficiencies in this. And now we only have to pay for three bids that we didn’t win, and all the labor and insistence that goes into that. You can see the geometrically progressive amplification of benefits just in this one example. There’s really-

(11:57) Yeah, go ahead, Eric.

(11:59) I’m sorry, I was gonna toss up some else out there though, is that one of the things that we see in here, it’s not just a potential improvement in hit rate, but that the accuracy of the improved hit rate is better. So it also is reduced time overall. So again, in that 20% hit rate, we’re assuming that maybe generally it takes equal amount of time to bid all five of those projects, if we’re improving our workflows, not only are we hitting those two out of five, but the time it took to produce those two is actually reduced. So we have, as you just pointed out, almost an exponential improvement in productivity in that application because of the impact to success in that area.

(12:41) Great insight, yes. And I like to remind people, it’s sort of here on the bottom, there’s really only three ways to to create value in an organization, generate more revenue, right? ‘Cause you can harvest value with more dollars. You can save money where you’re already spending it and you can avoid spending money that you otherwise would, right? These are all legitimate ways to create value in an organization. So as you’re identifying workflows, think about how you’re, which one of these categories you’re gonna generate value in or are you gonna generate in all three? So let’s flip to the next slide and we’ll get down to another insight. Look, this slide is essentially telling us the story that no matter how many times we try, we can’t get away from the fact that humans have to show up, leaders have to show up, managers have to show up and do management work. When I was a chief information officer, I was known for saying, my good technology cannot do your good management.

(13:42) And I’ll just leave that maxim with you to rest. Okay, the story is not gonna change. The technology tools are not the barrier. AI is advancing rapidly. It’s a valuable tool now. People are using and employing it, getting value out of it. Are there issues with it? Yes, they are improving. And those issues, whatever issues there are technically are actively being resolved. The adoption cycle is way, underway and people are adopting. The issues around this are organizational and the opportunities around this, if you’re gonna harvest value are organizational. We need to go from a distributed, everybody’s using the tools to an aggregated centralized management model. And that’s gonna come from strategy, understanding what you need. Of course, data protection and security concerns, but lack of understanding. We need to understand what workflows we’re working on things we’re trying to prove out and where we’re trying to harvest value, and then we need to train to it, and we need to put some honest money to it. So these are the old fashioned organizational capabilities that need to catch up. So this is a traditional capability maturity model where it goes from ad hoc and individual and moves up the ladder to defined and repeatable and organizational and team oriented. And that only happens with good old fashioned management control.

(15:04) And no matter how many survey responses we get, it’s proving this out. Okay, the next one. Structured learning 75% is self-taught. That’s not surprising, right? Because we’re still adolescent, we’re still adolescent with our use of AI because it’s mostly personal productivity. It’s edging into job productivity. It’s edging into team, work team productivity. And a minority of organizations are doing workflow, real meaningful workflow work right now. If we’re going to move up the maturity path, This basically is telling us that we need 75% of the people organizationally taught, which requires resources, structure, standardization, methodology, targeted use cases, accountability for outcomes. But if that’s the case, there’s only 10%, one in 10 are actually doing training, funding training doing training. So you can see there needs to be a flip here. This needs to flip.

(16:22) So, and on the bottom part, it’s just an error. I’m sorry. I did want to do something that this also tells me. And when we were looking at the results of the survey, and then I’m thinking about the one, the slide previous, the other finding, key finding is we talked about the fact that a lot of companies don’t have a clear strategy. What these things speak to is inconsistent results, right? So you have a lot of organizations are sticking their toe in the water and they may not be seeing the results that they think that they should. But it’s really no surprise. If we don’t have a clear strategy, we don’t know how we’re going to execute. We’re just kind of experimenting. And we are self-taught. We’re not investing in structured approach to developing the knowledge on how to use this.

(17:02) Yeah, so that’s a management function that needs to be sold.

(17:05) 100%.

(17:06) And that’s the hard work. I mean, the fun stuff in AI is already among us. Now the hard stuff is here.

(17:12) Yep, yep.

(17:13) Now the hard stuff is here. So, and what you’re talking about, we speak to in the very next slide, I believe. So let’s flip to that. By the way, Paul, if I could interject here, I should have caught this earlier, but in a prior slide, there was a comment that came in from Bill K. His words are margin is made in operations, not in estimating. I know you covered that, but I want to acknowledge the comment. Thank you, Bill or you could comment Yeah, it’s a So investment needs a plan. So do me a favor, Randall. Flip back to the previous slide, 75-11. Flip back to the next one, 77-11. We don’t have a dedicated budget because we don’t have a dedicated strategy because we don’t have a dedicated, we don’t have unity on what we’re going to do as an organization yet. So we’re not budgeting dollars yet, but we are spending money. This is going to do a disservice to organizations that don’t get their arms around it. You need to take that 77% of the money that organizations are spending. If your organization is in this pool, put a sponge on that, collect it, aggregate it, set some standards around it, create a real budget, align that budget to a defined strategy which says these specific workflows we’re gonna go and derive value from and then build structured programmatic training and development around that and proof, proof of value, test it out before you scale it up. So that’s really what this is saying. So Eric, we’re saying the same thing from two different directions.

(18:52) Yeah, 100%. One of the things that also jumps out to me is on a couple of these, we’re seeing the questions in the survey as well as the results say, here’s what’s currently occurring and here’s the direction we’re headed. What this speaks to then is a competitive landscape, is that ultimately we know the direction we’re headed to. Which organizations are taking those steps right now versus waiting and seeing what happens or maybe a little slow to get started on it. The competitive landscape is broad and it’s very diverse in terms of its application right now. It’s a tremendous opportunity if somebody can get their arms around this or for those organizations that get their arms around this improve business processes and then be able to implement those, they will be ahead of their competitors, right? And we’re seeing at 49% not having a strategy for implementing is, I mean, that’s a competitive landscape. I think that would be very opportune for a smart leader to take advantage of.

(19:52) Yeah, Eric, I’m just sitting here sort of imagining, do we have a football team? No, we’re not even in the game. Do we have a football team? Yes, we’re in the game. Do you have a winning football team? Well, that’s gonna require some coaching and some structure and some training and oh gee, okay, well that’s where we’re at. Everybody’s got a football team. Now we need coaches and training and real structure and organization or it’s not, they’re not gonna win. Right, ’cause everybody’s got a team now. Everybody’s got a football team. So are you gonna be a competitive football team or not? That’s where we’re at. That’s the All right. So that brings us making That’s right. And by the way, gentlemen, we’re making really good time, good pace so far. So over to you there. Yeah, good dialogue. So one of the things over the last month and a half, two months since our last webinars is we wanted to evaluate what are some of the highest and best use cases that people are utilizing right now.

(20:53) There’s, I’ve referenced this in the past that really this is like a kid in a candy store, you know, with 20 bucks. There are a lot of options. And we can take it a lot of different directions on how to benefit from AI. And I think that this is an exciting time because on one hand, you know, how do we narrow down into a strategy and take advantage of it? But on the other hand, we have this wide open space that says we’re only bound by our imagination. And so as we began to look at these use cases, where we wanted to take a look at what are being commonly implemented, which use cases are being commonly implemented throughout the entire project life cycle. Sure, we can set a project up in pre-construction and in the bidding phase to be maybe more accurate, more detailed, but then we still have to move through the project and execute all phases to be successful. Bill pointed that out. It’s important to understand that it’s the execution that makes the money. But in these use cases, wanted to cycle through a couple of thoughts on these different phases. Scoping and bid packaging a project up front are extremely important to your accuracy that will occur downstream. Evaluating the bid package as a GC, getting a set of drawings and getting specs from an owner then requires us to break that down and get those packages out to market so we can solicit you know, the RFPs from subcontractors and vendors. In doing that, this can be streamlined. It can be improved to take a little bit less time and be more accurate in what those packages look like. And then when those bids are received back, AI is easily used to evaluate those bids, evaluate what the coverage is, if everything that’s been requested is covered in their bids, what’s been excluded, what’s been included.

(22:39) But there’s also another piece here that’s fairly important is what areas of the bid of the subcontractor packages do we see overlaps? What tools do we see? Maybe machinery, maybe equipment, maybe requirements that are required by multiple subcontractors that a GC can then say, I’m going to provide this and eliminate that out of these multiple bids. And so it’s a way to, again, evaluate complete coverage of the bids, but also look for opportunities to reduce their costs as they move forward in a project and evaluate those with some intelligence. And in the past, those are things that took a lot of manual labor time to evaluate that. And now it’s just a matter of minutes with the right prompts. So it’s a really, this is a very good example where a lot of more enlightened companies I’d say, or maybe those with a strategy are starting to move this direction ’cause it’s a heavy lifting upfront with a lot of time at high risk.

(23:35) This is a good one. Moving into the next one, submittal logs and compliance, it begins in the pre-construction, but as we move through into the construction phase and throughout construction, we have to produce contractors, both subs, vendors, and general contractors produce a lot of submittals. So we want to know, number one, what are the submittals that are required? Do we have a full understanding of those? And what is expected out of those submittals? What is, what answers do those submittals need to be able to provide so that an owner feels confident that we’re going to do a good job with the project. And so identifying what that submittal, what the submittal list is, creating that log, looking at the specific requirements in there, and then even evaluating the information that we have coming from subcontractors and vendors. What do we have on hand that satisfies those submittals from perhaps previous projects? You know, some submittals are around, you know, for example, safety, what’s a job site safety plan? It’s not that difficult to use a previously submitted safety plan, make the adaptations for a project, and then there we have a submittal around a job site safety plan. And that’s a process that could be improved where just the timing to execute that can be improved. And again, the submittals that we have looking at the ones that are coming in, say from subcontractors and vendors.

(24:54) Are they compliant with requirements of the contract and in the specifications? So again, it’s one of these things that can be, that improves what was some heavy lifting and some very time consuming processes. And it helps also to raise awareness for those people who are using it to be aware of maybe where those gaps exist where maybe vendors are not providing the information that we really need and we might have to go back to them. So it improves our timeline and cycle time to review submittals. Change order entitlement, boy, this is another really good one. And this could take a lot of different looks. I mean, there’s a lot of reasons that a change order might be generated. It could be a requirement or requested on behalf of an owner. Or it could be a change order that’s simply necessary that we see something and it’s not really covered in the initial scope. And so it warrants a change. Now validating that change becomes incredibly important. Where in the contract was a particular specification identified or maybe not identified. So where did that occur? Did we have guidance anywhere else in there indicating that the work would be done? one way or the other, or maybe it could be a product from a particular vendor. And so gathering that evidence, it becomes important. And then building the information that we need to submit a change, whether it was a directed change that came from the owner, we still have requirements that we need to comply with for that change to be approved. If it wasn’t a directed change, it was a constructive change, one that was necessary in the course of construction we have to build a better case for that. And so that constructive change is going to warrant a little bit more deep dive on documents, what’s authorized, what was specified, et cetera. And so again, this is something that can be a lot, can be streamlined and the accuracy improved using AI and using some deep search capabilities. All of this in the context of what is required in the contract. If we get an opportunity to then present our change and it’s more, we sit on solid ground, it’s more valid in the eyes of an owner, well then we’re more likely to collect the value of that, right, and get paid. It’s not unusual. Change orders negotiated and priced at the end of a job typically occur at a smaller percentage of their value than if they’re negotiated and collected during the project. But oftentimes the heavy lifting of pricing, working through entitlements on changes, costs us to shift that to the backside of the job. Also not unusual to group change orders together, then they get negotiated at a smaller price.

(27:35) If a change order is thoroughly researched in real time, and we make a better case for collecting full value on that. So there’s real money that’s tied up in just effective change order management. Of course, contractors know that, but using AI to help focus on that entitlement component can become extremely important and improves cash flow cycles throughout the project.

(27:59) Yeah. Eric, what you just described here is deeply, this is the devil is in the details in change order entitlement, right? This is not rocket science. This is compliance, details, lots of analysis and research and crossing T’s and dotting I’s, right? It’s all the minutiae. We have to do the minutia faster and we have to do it with much more accuracy. And that’s where the opportunity is. So highly labor intensive. A lot of that gets pushed.

(28:31) The other piece of this is that you can actually leverage it to get you more accurate cost information on what the impact of a change. You know, oftentimes we think about, you know, what’s the cost of executing a change? Well, typically materials and then we have labor and we have potentially equipment impacts. But if we really deploy AI to do a little bit deeper search on what’s the potential impact of this change, we start to look at the schedule of re-sequencing. It might be shifting of coordination of subcontractors, other trades. Those things add time and complexity to the project that is very underappreciated for most owners. And so by doing a little bit deeper research, you can make a solid case for extended overhead or delays in a project based on a change and give a lot better case for asking for a schedule extensions if that’s the case.

(29:24) You’re describing a scenario where slowly adding points to margin, you know, 10th of a point. Just a little bit more margin and a little bit more efficiency, running a business a little bit better, that’s a leverage point.

(29:38) And for most contractors, it may not be adding points, it’s just not giving them away, right? That’s exactly the deal is they’re out the, you know, margin fade occurs for a reason and good tracking of change orders is an excellent way to–

(29:51) That’s a very good characterization.

(29:53) Yeah, I mean, let’s just worry about keeping the ground we have before we talk about getting ground. Let’s just not lose ground. That’s a good first goal.

(30:04) Yeah, it’s an excellent–

(30:05) Sold margin versus delivered margin. I don’t know how many contractors really view that. If Bill were here, he would love to comment on that.

(30:14) You know, just broadly, broad use, I’ve talked a little bit about it, but just doing deep search in your documents is, you know, I would characterize this as one of the low hanging fruits of use cases right now. How do we extract information during a project typically before AI? Well, it’s flipping pages, looking through specs, highlighters, red pens, those sorts of things, right, to locate items that are in the contract or the specs. So this just streamlines it, and it could be a variety of different things. It may fit what we were talking about in the previous one about change order entitlement. It might be validating specification. It might be maybe needed to hold a subcontractor accountable or a vendor accountable for what was procured. So gosh, there’s a variety of different things. It also can help mitigate the impact of owner requested changes. Sometimes we see that owners request things that might not be picked up as a change.

(31:10) And so by documenting what’s in the project versus, or what’s in the documents versus what’s just simply an ask can help clarify that. Verifying specifications as we have in the example here with doors, but it’s a million different ways. This is, I think, the biggest opportunity for contractors to stick their toe in the water, so to speak, for use cases is just search your existing documents for those things that are pain points and you’ll begin to find all kinds of results. So yeah. And then moving on to our last one. Yeah, close out and cash collection. So this is another such an important one. You know, we talked about revenue, we talk about managing our costs and then maximizing our margins. And this occurs really in that last, the closeout process, ensuring that we’re complete with closeout requirements, documents, as built, you know, whatever that is warranties, lien releases, a variety of different things that might be necessary in the closeout process. Closing out tight and efficient allows us to collect final payment and retainage. This is extremely important. The margins in this industry are not huge. So anything we can do to improve the cash flow cycles for contracting and vendors is going to help the industry overall. And so when we are tighten up the closeout process, aggregate documents, scan for requirements that we might not have missed for closeout, those sorts of things, we improve that cash collection cycle. And that adds solidly to the bottom line. It is very It’s very frustrating to watch a lot of contractors that complete the job in the schedule, but they get to that 95% point and that last 5% starts taking an inordinate amount of time to extend out to get that final payment. So this is also a very important opportunity to use AI to just improve document management and check boxes and just make sure that we’re following those close out and turnover requirements as accurately as possible, as thoroughly as possible so that we can finish that project, move on to the next one. It’s also, there’s another piece in here about how work that occurs under the main contract versus warranty you know, that sort of break point in between there. So if we manage to close out a project efficiently, then we can move into a post construction and not give away so much work.

(33:44) Those are some example use cases. I’m engaging with clients in the industry all the time. And these are the items that I’m seeing that people are wanting to use. They are using it to various levels of sophistication. But as I mentioned before, all these use cases speak to this is a very competitive landscape. And wherever we sit on that continuum understand that competitors are out there looking at these tools to try and improve their workflow processes.

(34:15) Yeah, Eric, I would totally agree. It’s safe to say that people are actively engaging use cases now and they’re moving away from personal productivity and they’re figuring out what use cases, what workflows are we gonna go harvest? first in organizing around those. So this very much is a competitive opportunity. I see that. The other thing I wanted to add to the use cases conversation is just a reminder that each one of these use cases that Eric went through are included in a follow-up document that we’re going to share with all the attendees here today. You can download from the website later as well. And it’s 20 use cases, including these five, including prompts and suggestions to get started and things to think about. So it’s a very rich document that you can take for your team and get your heads around that, get your team around that and start testing stuff out. So that should be a useful tool for everybody. I just want to remind them that’s available to them after this.

(35:16) That’s a good call out Paul and Eric. And by the way, I’m monitoring our Q and A. I don’t see any comments or questions yet. I’m gonna dare suggest that because I know our final two slides are so content rich. Eric, Paul, let’s let her breathe. All right So the In this one, what we were looking at, Paul and I began talking about this, gosh, I wanted to say six months ago now, realizing that there is a logical flow. And what we’re seeing is, and you already articulated, Paul, is that we’re moving from personal productivity, which is really the me and the my job, into a time when people are looking at their workflows. That’s again where this competitive landscape really begins to take form is what are we doing to improve our workflows within our teams and within our departments, et cetera. And how does that get better? How do we share information? How do we learn consistently with each other? That’s becoming a what we’re observing is that that’s becoming extremely important and influential in how successful AI is for the companies that are deploying it. This is not, it doesn’t mean that we guarantee success or don’t guarantee success if we’re following this path, but 100% we’re observing that the maturity follows this process pretty cleanly. And the interest right now is we saw in the survey results is indicating that there is an expectation of getting to that level four in here in this workflow is that we’re improving processes, we’re improving processes that have real impacts on the bottom line.

(37:03) I would say, Eric, you and I had agreed in the pre-meeting that if we could rebuild this slide, we would emphasize number two also.

(37:13) Sure.

(37:14) And I think if I just go back to my football analogy, if you guys will suffer through that with me. Look, we’ve recruited the team. Now the team needs to work together and we need to go take the field. We need to win the game. And, you know, that’s really no different than any maturity model in any scenario, you know, of life. We need to apply it here. I think it gets obscured because it’s cool technology, but we have to go back to the fundamentals and that’s what these are. And even when you’re on a team, you have to play your position well. That’s what box two is. That’s what circle two is. You have to do your job well. What is your contribution to the team with the AI readiness? And what is your portion of the workflow that you’re activating and working on? And so it’s what are you doing individually? What are you doing collectively? And qualifying yourselves individually in increasing your own capabilities and then increasing your team capability. This is a classic maturity model. It’s just being applied in a new environment, but this new environment is probably moving faster than any other technology environment has and it’s so accessible, we’re saturated in it, but it’s almost like being saturated in chaos. This maturity cycle will bring structure and value. It’ll aggregate what you wanna do. It’s like looking through a prism and focusing the light on the other side.

(38:48) Right? Like a magnifying glass, collecting all the light that’s all over the place, focusing it so that you can harness the energy and the value out of it. So it’s time for leadership to step up to the plate. That’s what this says. And 2027 is the year for that. Paul, I’m thinking real quickly back on the work that we’ve done in the past together. And one of the things we always discovered is that a lot of tools have to get pushed out, especially around technology, because people are comfortable in their zone. They’re comfortable in what they know. Yeah, they’re comfortable doing it the old way. They are. And in fact, that’s a common argument. And I just like to do it my way, right? And that’s a reasonable argument because that speaks to confidence in the system that you can control, right? And then as you move into new technologies, people can be resistant to that. What we’re seeing here is the opposite.

(39:38) Absolutely opposite.

(39:40) This is an organic pull towards using technologies and using systems that have a phenomenal

… and what we’re seeing is

organizations that are saying, hang on, we’re not sure what to do with this just yet, let’s figure out a logical path forward

in a rich environment, and ’27 is going to be a big year in an environment where

They’re ready. What they need is leadership. Let’s guide this readiness so that the deployment produces the strategic advantages that we’re looking for. But it’s interesting because it’s different in the past where you have a flight. I think there’s a management sentiment out there that’s floating around. I don’t want to dampen the momentum. If I put too many rules in place, if I make too many standards, if I assert too much management authority, we’re gonna dampen momentum. I think that’s, I’ll just say it bluntly, I think that’s a cop out. We need to get through this cycle and we need to get through it sooner than later. Let’s put some structure in place because people actually will thrive in a structured environment where they know what to do and what’s expected. And equally, what’s not acceptable. Yep. Right. Yep. So, all right. You want to grip on this one a little bit? Yeah, we have a lot of this up. And so this is kind of what we see as the moves. This is what leadership really needs to be thinking about right now.

(41:07) First of all, who’s taking [ownership?] of this inside the organization?

What we’re seeing is kind of

right now. It’s at a very

as we’re paying for licenses, as we’re beginning to develop

Who owns that? Is it an individual? What does that look like?

now has, or is going to have, or what is, you know, what tools are we gonna use? How are we gonna use it? What are the controls, et cetera? And so we wanna begin to identify who’s gonna be the driving force in our organizations right now. We see mixed, you know, the surveys shared that we see this mixed, you know, one of the certain, one of the things that jumped out is, you know, use at the executive level is high but at the same time, it lacks leadership in a structured manner. So this is extremely important. By the way, I don’t feel like the senior executive of an organization needs to be tagged as the owner of this, right? It’s going to need some focus, right? It doesn’t mean that the executive is not engaged in it, but there’s a lot of moving parts that needs a lot of attention.

(42:13) The controls are pretty straightforward. We need to define how we’re using it, what’s okay to do, what’s not okay to do, which tools, which information, how to protect that information. This becomes extremely important. I think that’s going to feed the next research that we’re doing, we’re engaged in right now. By the way, quick teaser coming up. We are engaged in research right now about risks and legal implications about AI use. And we need to be paying attention to this strategically as an organization. So we will impact more.

(42:41) The risk component of that, we’re taking a deep dive into the insurance industry. How are they reacting to it?

(42:47) How are they reacting? How are they treating the risks? How are they treating events that are influenced by AI? Very important. And yeah, we’ll come back on that one. One of the workflows that we wanna use, that’s the next thing that a leader or a group, an owner, if you will, needs to determine is, What are the workflows we’re gonna test to figure out how do we use it? How are we gonna maximize value? I think this is great. This is where we might look at what’s occurring during pre-construction estimating. Maybe it’s can we improve how we manage certain steps during construction? Closeout is another one. But the point here is you pick a workflow and you begin to measure it. How do we do it now? How do we improve it? What are the results? Did that create the impact? And that gives us a sense of control and strategic deployment of the technology. Capability then begins to look at a little bit broader picture. We wanna make sure that whatever we’re using whatever we’re deploying that we’re training people appropriately to use this. And one of them, I didn’t mention it before, but what jumped out to me is just as basic as prompts, using consistent prompts in a workflow to generate consistent results. If I have three people in estimating and we’re all prompting differently, we’re gonna produce different results.

(44:03) So that’s part of that training. Is it what tools? How do we use it? How do I prompt the right responses? How do I validate the information I’m getting out? So we want to build that specific capability in those workflows. And then finally, the last piece, which is, you know, any senior executive, this is what they’re going to be paying attention to. What’s the value? What’s the risk? Did I reduce risk? Did I increase revenue? Did I increase margin? Did I reduce cost? Did I reduce risk? All of those things are critical. We need to track and measure it. Right now we still have somewhat of unstructured deployment where I don’t believe we’re maximizing value. We’re still in the gee whiz stage of this. This is great look, it can change you know, my shirt for me, right? Everybody’s excited because I can now create head shots without having to find a photographer, but how much real value is that creating? So I think that that’s the next major step for leadership is to say, let’s use these tools where we’re maximizing value, reducing risk, et cetera.

(44:59) Yeah, I would just say, I would add on this, you know, as I think about number five, measuring, you know, value and risk. You know, how do you do a business case on this? That often comes up, you know, in larger organizations, a program management office is designed to do this with business analysts, time motion studies, understanding rework, understanding transactional volume and where the hits and misses are in that quality and qualitative work, not just quantity of work cycle times, all real dollars being spent, soft dollars being spent, how do you measure productivity? Business analysts are good at this, right? In larger organizations, smaller organizations need to rent that talent to help you get started. Go to your start with your CFO, start with your accounting team right, but you have to give this an honest effort and put a lens on this. I strongly advise to start exercising your own organizational, developing your organizational capability to become [unclear] independent. You need to know how to measure your business. And if you don’t have a PMO, you don’t need a PMO, but you need a little bit of help.

(46:13) You know, this is such a good example is that small firms can become highly competitive with big firms because they are now able to examine the entire workflow process with more clarity and understanding impact of that workflow process. So I think that there’s a potential leveling of a competitive playing field out there using the football team analogy.

(46:38) I could not agree more. This literally is the definition about outmaneuvering Goliath. This is my club. It’s the middle markets moment, right? It’s the SMB moment to level the playing field and have access to the same tools. Yeah, very good. Yeah, that’s good insight. So really, the good news is AI is among us. The other news is the future is in your hands. What are you going to choose to do? And I think that the, just speaking of what we’ve been talking about, a little bit of a common thread in here is that it is here, it’s impacting, and I think we’re proving the fact that it’s not gonna go anywhere. Now it is to maybe act with a sense of purpose and urgency. And I don’t, and I’m going to start with purpose first because that speaks to strategy and leadership, right? But to be diligent and execute on purpose, not ad hoc. And I think up until now we’re seeing that landscape. That’s what the research to date has shown us is that, yes, it’s exciting and people are adopting, but it is not led. It’s not strategic in nature, but now is the time become strategic, we’ll map out what does it look like over the next couple of months? What do you and I think I’d say next couple of years, but that is almost behind, right? We really need to be looking at very near term. What are you doing right now to advance because and I referenced this when I’m working with clients, there are very smart people sitting in rooms just like this planning how they’re gonna beat you on the playing field. And so everybody has, there’s a sense of urgency to don’t let this get stale. Now is a huge opportunity to get in front of your competitors, define improved processes and execute with more efficiency and ultimately profitability, which is ultimately the goal. So this next 90 days is not just tongue in cheek, it is very critical. Everybody needs to be acting.

(48:38) Training and fitness regimen. Good old fashioned discipline, focus and discipline. Set a goal and carve out two or three workflows and go prove them out.

(48:49) Yep, you gotta start.

(48:50) You gotta start.

(48:51) Yep, you gotta start. We have fewer than 15 minutes left in our session today. I wanna make sure that we cover off Q&A. We did manage to gather a few questions from our community ahead of time and during our session. I’ll tee up the first one, but I also will share a reminder and encouragement for our audience members and participants. If you have a question for Eric and/or Paul, please feel free to submit it. There’s two ways you can do that. Use the Q&A function on the Zoom ribbon or dashboard. Or if you’re willing, put up your hand. Use the raise hand functionality on Zoom. Our back-of-house team will let you into the room, if you will. And all I ask is that you come off mute, come off video and you can have a one-on-one conversation with our speakers today. But here’s the very first question that we were able to gather through this session. And I don’t know who is best to answer it, but I’ll ask it. Which use cases would you call low risk enough to pilot first before the governance is fully in place?

(49:54) I don’t mind taking a first stab at that.

(49:56) Go for it.

(49:57) Yeah, go ahead, Eric.

(49:59) So, a piloting use case that you would characterize as low risk so that we can kind of experiment with it, develop some familiarity. I would say looking at deep, deep search and research with existing, with your documents, both internally and externally. The reason I say that’s low risk is because you’re not really committing to anything. You know, there’s a lot of work that we do on a regular basis that, you know, in terms of research or gathering information, that’s labor intensive. It takes a lot of time, right? And so we can experiment with ways to find certain things, whether it’s in our contract documents or a company files, project files, et cetera and be able to identify information that before may have taken more time. And so it really helps. This is not too different than opening up a document and searching for a word, right? We do this, we kind of have a familiarity with it already. So this is just using it a little bit more sophisticated level across a set of files, a set of documents. It’s low risk because And for some reason, there’s something that’s not identified there. We missed something. We’re not affecting, we’re not changing anything. It’s just getting familiar with the capabilities. What it will also allow us to do is get better at how we prompt that search and what we look for specifically. That takes practice. Nobody’s gonna be good at this coming out of the gate. So it’s important for those deploying it to spend some time searching over and over until you begin to refine, ah, this is how I find the exact answer that I’m looking for. And now you have something that you can replicate actually into workflows. So I would say that that’s a lowest risk one because it’s not tied to anything directly that’s financial or that might get you in trouble.

(51:38) Yeah, Eric, I have a little bit different view from a different viewpoint on this. I think there’s low risk and then there’s managing the risk that you have that you’ve identified. And that’s sort of what we were talking about at the beginning. I don’t want to miss this point. That part of the opportunity, if you just take it like estimating that we gave as an example is. It’s an elevated risk activity in the industry and it requires a lot of human labor and intensity to mitigate that risk. What’s the risk? Accuracy. We need to be accurate and we need to be as right as possible. We need to be as fast as possible. And we need to be as competitive as possible with the right truth. AI enables that by creating a research assistant. So, you know, we’re managing our risk more effectively so that it’s not borne, right? So one way to avoid risk is to go towards low risk opportunities. Another way to avoid risk is to manage the risk you have better.

(52:49) And that’s what I think is happening with AI and just if we just keep going on our estimating examples, you’ve got a lot of deep research and analysis and data synthesis and cross checking and double checking, et cetera, so that the humans who are in the loop now have the opportunity to exercise judgment, look for accuracy, make better informed decisions faster. Now that’s managing the same risk you have better and faster and cheaper. So that was just a different perspective that’s taking high risk and making it lower risk.

(53:24) Yep, yep.

(53:25) You know? So that’s what insurance wants. They want to convince you it’s a big risk and then they want to push you into managing that risk so they have less exposure. But they want you to pay for high risk. I’m going to apologize to the community because there’s no way we’re going to get through all the questions, but I am able to tee up at least one more. So forgive me if you did have a question for Eric and Paul. Rest assured there is an opportunity for you to connect one-on-one or with our two speakers today. More on that very shortly, but here’s our one and only our final question that we have time for, and that is, and I’m gonna paraphrase this, we don’t have an AI policy yet. So if you were writing version one this week, what are the three or four things it has to cover?

(54:20) Okay. You know, I could think of a couple of things right off the top. And again, this is interesting because there are, you know, some organizations are beginning to develop their policies, but a lot of them don’t because obviously everything that we’re talking about, it’s very organic. That being said, there’s, there’s, I would, I would put it, put policies need to fall into three categories. It’s just, right off the top of my head. The first is security and confidentiality. So how do you make sure that the data that you’re using and the analysis being conducted is secure? So we need to be thinking about what do we allow to be uploaded? What are we using AI to access and how do we know that it is secure? So I think that security piece needs to be looked at. Which tools are we using? What information is being used in those tools? And having some guidelines around there. One of the key things that needs to be very upfront is that we should not be using information that is explicitly protected via contracts or confidentiality. The reason is because I was just reading something and we’ll unpack this later, but that AI now and the conversation that are being used is discoverable, right? It’s discoverable and it’s potentially already out at risk on the internet. So we have this thing that we want to use it for business process and reviewing say contracts and things of that nature. But if we’re using it and we lose sight of the security aspect of it, we open up some pretty substantial risk there. So security is one. The second is behaviors. What should we be using it for? and what should we not be using it for? And putting some parameters around that. That speaks to getting some standards in place. What’s it OK to use and what’s it not? And it could be also speaking to business processes. The last is quality of the output.

(56:13) AI does not replace your intelligence. It does not replace your personal experience and the knowledge that everybody has. So we have to make sure that the output that comes from AI, and if we’re asking it to review contracts, identify specs, or identify gaps, or whatever, is that we’re smart enough to be able to evaluate whether it gave us the truth because we can’t make business decisions on bad information, we can do that now. We don’t need another tool to complicate it, right? So we need to have some protocols that ensure that there’s quality behind the work that uses AI that uses AI.

(56:48) Yeah, I might add to that a little bit. Eric, I think that was you covered the gamut there. In my simple mind, I think standards are imperative. What tools are we using? What tools are we not using? Procurement, how to procure, what’s acceptable in procurement? Acceptable use, which goes to your sort of intellectual property and protection and security what’s acceptable use and governance, you know, around this. So it’s the rules of engagement for us that we’re going to agree to as an organization. So, yeah, those are the fundamentals.

(57:20) Excellent. I’m gonna take this opportunity to, yeah, it was. I’m gonna take this opportunity now to draw to a close, make sure we end on time. Big thank you to our audience members, certainly to our speakers, but there are multiple resources available to you. And so I just sort of quickly recap those. Copy of the AI Readiness Report will be distributed afterwards. Certainly, as Paul had mentioned, there’s 20 use cases of AI. The recording to this live event, this webinar, will be shared out as well. And as I teed off earlier, there’s an opportunity for you to sign up for your very own AI Readiness Check-in that’s available for everyone here. If you’d like to schedule a one-on-one or a meeting with either Eric and/or Paul. That’s available to you. In fact, our back of house team is dropping. That URL right now for you to log in and sign up, it’s frictionless. We just need your name, information, name, email, and phone number. That’s right. Yeah, absolutely. And as I plugged earlier, our next installment of this virtual series is going to deep dive into risks. So be on the watch out for more details in the weeks to come.

(58:24) And Eric, we can’t end this call without saying a big thank you to you for your collaboration on behalf of TAG CXO. We’d love to just toss to you for executive summary and any final calls to action.

(58:38) Sure, yeah, thank you. Thanks Randall, I appreciate it. This is such important information and we’re in this gathering phase. And I actually call to action for me would be related to the next one, which is that we are in the middle of sort of deep dive on the risks and legal implications. And literally last night I was reading an article about some legal implications about how AI is being used we’re beginning to start to, we’re at the front side of seeing some case law begin to develop. This is incredibly important. And I certainly feel like we’re in a unique position to be able to do research and bringing it to the table saying, don’t make these mistakes. This is what maybe other people might be stubbing their toe on, but let’s learn from this. So. What I want the call to action, one of these calls to action is if there is, if you have concerns, if you have risk or legal concerns about AI is being deployed or maybe deployed or. However, the risks, the mistakes we may make.

(59:38) Ask us about it, engage with us. Say here, I’m concerned about this. What’s being done? I don’t know that we have an answer, but it allows us to ask that question in the research because we’re going to go back to the industry, ask more questions. We’re going to engage legal experts, insurance experts to understand what’s the demeanor right now and what information can we gather and share with the industry that are going to be helpful. Ask those questions of us and we will go seek them out as well. We’re gonna use our network to try and address some of these questions and bring answers to the table.

(1:00:07) Well said, Eric, excellent.

(1:00:09) I’ll speak for me and Paul and the rest of the TAG team. Thank you, Eric, and for everybody else, we will see you on the next one. Bye for now. Enjoy the rest of your week.

(1:00:17) Thank you.