Latest / Engineering Influence from ACEC / From Curiosity to Impact: How AI Is Transforming Engineering Firms
Transcript
- 0:05Welcome to the Engineering Influence Podcast from the American Council of Engineering Companies.
- 0:10And today, we're going to be diving into a subject that is of great interest
- 0:14to our members, increasingly so, and it's actually going to be a big focus of
- 0:18our fall conference coming up at this point in this recording next month in October.
- 0:22And that is AI. And not just AI adoption,
- 0:26but doing so as a strategic lever to improve a business and to make things work
- 0:33more efficiently and to discover new ways to build business.
- 0:38And to do so, we're joined by two people who can speak to that on the provider
- 0:43side and then also the user side.
- 0:44Very pleased to be joined by Tanya Kennedy-Flood, Corporate Marketing Communications
- 0:49Leader at RJC Engineering. And then also Eric Liu, he's the product officer,
- 0:55chief product officer at Open Asset.
- 0:57And I want to welcome both of you to the program today. Thanks for joining us.
- 1:03Glad to be here. looking forward to it. Likewise, thanks for having us.
- 1:07Yeah, it's a great subject and something that we really want to delve into,
- 1:11especially since we're away from the whole idea of the new shiny thing and we're actually using AI.
- 1:18So Tanya, let's start off with you and give us a little background on the firm,
- 1:23your size, kind of location, what you do, and kind of how you were approaching
- 1:26AI from the business side.
- 1:29Yeah, absolutely. So I'm with RJC Engineers.
- 1:33We are a North American-based engineering firm with offices across Canada and into the United States.
- 1:39We focus on structural engineering, everything from new design to restoration,
- 1:45building enclosures like high-end glass design or building science like the
- 1:49envelope, etc., and do work globally as well.
- 1:52So, as you can imagine, with 700 RJCers, we're increasingly looking for efficiencies
- 1:58and ways to be able to better deliver our services to our clients and certainly
- 2:04see that opportunity in AI.
- 2:07We, though, I think we started like many firms in this space and,
- 2:12you know, fairly many businesses with a little bit of hesitation.
- 2:16So we knew that the potential was there, we could see it,
- 2:20but we wanted to make sure that we approached AI in a really thoughtful and
- 2:24intentional way, that we will provide them with enabled workflows and tools
- 2:29and recognition that AI is not a fad.
- 2:33And there's a big change when that came that's really fundamental from what
- 2:38I'm going to call like traditional AI to more generative and perceptive AI,
- 2:42that fundamental shift to tools that we really can use in the AC space to be
- 2:48very efficient and effective.
- 2:50So we as a firm made an intentional investment.
- 2:54We brought in a vice president of business technology.
- 2:58So someone with advanced expertise in this space that's forward thinking and
- 3:03able to help lead us on this journey to help us understand, you know,
- 3:07what's possible, how to implement AI responsibly.
- 3:11From there, we've put together strong foundations. So,
- 3:14and I would recommend this to everyone, an AI policy to give clarity to every
- 3:19RJCer on what is appropriate use of AI and an AI working group to help guide
- 3:25and facilitate adoption across the firm.
- 3:27So having those in place allows people to be encouraged and to experiment,
- 3:33but it also gives us some safety.
- 3:35And it's been really key, quite honestly, to building the momentum within the
- 3:39firm. So we have started off small with measurable pilots.
- 3:45So things like, you know, Microsoft Copilot or OpenAsset, where people can see
- 3:50the immediate value in their day-to-day work.
- 3:53Those existing systems that have AI capabilities, I think that's a really great
- 3:57place for firms to start.
- 3:59You know, our journey, it's really been about moving from curiosity to actually
- 4:05like structured exploration.
- 4:07So piloting these projects, improving workflows, not rushing in,
- 4:11but being very intentional and setting ourselves up on a thoughtful path that
- 4:16ensures AI will actually solve our business challenges and not create them. Yeah.
- 4:23There are a couple of points there that are great to dig into.
- 4:26Before we do that, I want to go to Eric and kind of get your perspective from
- 4:30the open asset side, you know, with a firm like Tanya's going into like exploring
- 4:36this and getting this actually set,
- 4:38you know, how did you work with RJC to help them on their journey and selecting
- 4:45the best product possible to get them started?
- 4:49Yeah, that's a great question. I mean, honestly, for us being that we are a
- 4:52tech company and we are a product company, It's we help RJC as well as all of
- 4:57our other customers by just providing a really great product.
- 5:00At the end of the day, if we focus on the product values of how we leverage
- 5:06new technology like AI to solve these pain points within the AAC industry and
- 5:10leverage it in a way that's intuitive and meaningful and safe and secure,
- 5:15it's really leveraging that new technology to solve these pain points in a very unique way.
- 5:20So obviously, with our open asset product, that is an AEC marketing platform
- 5:25that really helps AEC firms with digital asset management, as well as winning
- 5:30new business from a proposal standpoint.
- 5:32It's really about really understanding the nuances of the industry and how we
- 5:37leverage AI in meaningful ways to help solve these problems.
- 5:40So that's kind of what we do specifically, not only just for RGC,
- 5:44but really for all of our customers as well as our prospects on making sure
- 5:49that we're building really great solutions that actually solve a pain point
- 5:53in a very meaningful way.
- 5:55So let's dig in a little bit, because that's great, you know, Eric,
- 5:59on, you know, the asset management, digital asset management,
- 6:03which is a massive undertaking logistical challenge for a lot of firms and then
- 6:07firms that have a lot of history and a lot of RFP responses.
- 6:12And you've got a wealth of information,
- 6:16really approaching that and trying to figure out how we're going to add that,
- 6:19you know, you're going to add that into a model and actually use that for,
- 6:23you know, not just another tool, but, you know, just something that can actually,
- 6:29help create efficiency.
- 6:31So, Tanya, how did the firm...
- 6:33Kind of prioritize or identify things that they wanted to actually apply with
- 6:38AI technology? How did you guys kind of start that process?
- 6:42Well, you know, it really resonates with me when you say history.
- 6:45We are a 77-year-old firm.
- 6:48So we've got a ton of data, a ton of assets, et cetera.
- 6:54So yeah, it can be hard to know where do I even start, right?
- 6:58And sometimes it's starting by learning.
- 7:01Like when Eric was talking about solving problems, like one thing,
- 7:04for example, that was an early implementation of AI for us was an open asset.
- 7:10You can search for images using native language. So I could put in like bridge
- 7:14at sunset and now it goes through all of our images and we'll bring up bridges
- 7:19at sunset, whereas before you'd have to manually tag them, et cetera.
- 7:24So, you know, going back to the question, you know, deciding what to automate,
- 7:28it's not really about chasing a tool.
- 7:31It's really about solving your real business challenges, right?
- 7:34So recognizing like, what is it that's kind of painful that we know we could improve upon, right?
- 7:40In the AAC industry, there are so many opportunities.
- 7:44Like research suggests that I believe it's 30% of current tasks could be automated by 2030.
- 7:51I personally think that's going to happen even earlier than that,
- 7:55the value comes from asking, you know, like, which of those tasks actually matter
- 8:00to our people, and ultimately to our clients, right?
- 8:04So looking at user workflows, that was a starting point for us, right?
- 8:09So automating what we already do, focusing on tools that fit into our existing
- 8:15processes, rather than just, you know, automation or AI for the sake of it,
- 8:19but it allows people can immediately see how it saves them time or how it benefits them.
- 8:25Having those tools that
- 8:28you were using or those workflows and putting things
- 8:31in place like it's really a seamless and easy way
- 8:34for people to explore and use ai and speaking of our people deciding what to
- 8:40automate we ask their people like we involve them they're the people ones that
- 8:45know where the bottlenecks are where the opportunities are in their day-to-day
- 8:48work so we surveyed people and said,
- 8:52you know, where could AI make your life easier?
- 8:54What opportunities do you think our firm should be capitalizing on?
- 8:58So that we were addressing meaningful problems rather than just imposing what
- 9:03we think is right from the top down.
- 9:06And then having those guardrails with that AI working group.
- 9:08So all ideas come through that group so that we can prioritize based on business
- 9:14problems and efficiency gains, integrations, as I mentioned, with existing systems.
- 9:18So really looking at what's the best balance between innovation and user value
- 9:25and ultimately, you know, what we deliver to our clients. So that's what's going
- 9:29to make AI usable and meaningful long-term for any firm.
- 9:33So I want to extend all that and bring Eric in on this as well,
- 9:38because, you know, Tanya, I want to hear about what you heard from that AI working
- 9:44group on what What the pain points were and what hopeful solutions AI could help bring about.
- 9:50And then, Eric, on your end, you know, working with clients,
- 9:55you know, how do you guide them through that process?
- 9:57I mean, a lot of the times it's, hey, how can AI fix this or how can we use this technology to...
- 10:05Make our lives easier. And you don't really know where to start because it's kind of a blank page.
- 10:12So Tani, starting off, I mean, what did you guys hear from that working group
- 10:16and then Eric and, you know, how do you help clients like RJC kind of throughout that journey?
- 10:22Yeah, we heard everything. So all of the surveys went to all 700 people and
- 10:30then came back through the AI working group.
- 10:33And it was fascinating to see the commonalities, right?
- 10:37Everything from minute ticking, for example, super easy, low hanging fruit for
- 10:41AI to, you know, in Eric's world, proposals, right?
- 10:46Like so much, we are a sub-consultant and we're also a pride consultant.
- 10:50And there's a lot of opportunity for AI to bring about great efficiencies and
- 10:57insights, quite honestly, for proposals.
- 11:00We heard everything too, though, from like what we could do on a client side
- 11:04with a parking garage, for example,
- 11:07and how we could utilize the data we have from years of designing these structures
- 11:12and repairing and maintaining these structures
- 11:14to bring real data-driven decision-making powers to our clients.
- 11:19So it was really quite interesting and glad that we have a really well-rounded
- 11:25group to help filter through to make decisions on where do we start and what
- 11:30should those priorities be?
- 11:32Yeah, I think the way that Tanya and her working group at RJC approached this
- 11:37problem is really the ideal way.
- 11:40Like she primarily focused on what are the pain points that we're trying to
- 11:43solve and then really got feedback and survey research on some of the things
- 11:48and opportunities within the company that potentially could be benefited from automation via AI,
- 11:54which is honestly just really the proper way to do it.
- 11:57It's really no different than how we do it as a tech company from a product
- 12:01perspective, really just focusing on the customer problem and the pain point
- 12:05first, and then seeing how we leverage new technology to solve it.
- 12:10AI these days is so amazing and innovative that a lot of tech companies out
- 12:15there are kind of starting with
- 12:17the technology of AI and seeing where to apply it to solve pain points,
- 12:23which honestly is really a backwards way of doing product development because
- 12:27you're basically just looking at the technology and figuring out how do I shoehorn
- 12:31it and just give it the label of AI as a marketing gimmick when you're not really
- 12:37necessarily solving these pain points that are really worth solving.
- 12:41So the way that Tanya approached it with RJC is exactly how we approach things
- 12:45from a product company perspective is really truly understanding pain points
- 12:49within an organization and then seeing where AI can potentially help with that.
- 12:54In the realm of automation, part of our being able to apply it well is understanding
- 13:02the technology as well and figuring out, is AI good at this or not so good at this?
- 13:08Because I think in this overhyped world where AI is seen as this magic silver
- 13:13bullet that pretty much can be applied to anything and everything and be really
- 13:16good at it, that's actually not really the case.
- 13:18There are some areas where AI is really, really effective in,
- 13:21like automation, like content generation, like conversational UI.
- 13:25But there's also a whole bunch of things that AI is not quite there yet in terms
- 13:30of being really, really effective, which is around prediction and forecasting
- 13:34or planning and strategy or truly autonomous systems that require zero human interaction.
- 13:40So I think understanding the technology, what it's good at as well as its limitations
- 13:45is also something that obviously is very, very key for me and OpenAsset.
- 13:49It's just making sure that we're applying the right technology in the right
- 13:53moments to solve actual use cases.
- 13:56Yeah, I agree with you 100%. I find that in my work in, you know,
- 14:01in Tanya's world and in kind of the marketing and communication side,
- 14:04it seems like every product has something shoehorned into it to be AI or,
- 14:09hey, here's our new AI feature.
- 14:11And they rush to market and sometimes it's half-baked, right?
- 14:15And usually they're trying to apply the technology without much direction, right?
- 14:20And it's important to understand exactly why are you using this and whether
- 14:23it's a good fit or not. But, you know, it's kind of that the shiny object kind
- 14:29of trap that you fall into, right, that it's brand new and you want to apply it.
- 14:34And then you realize what the limitations are for, you know,
- 14:38responding to opportunities for new business, for going back and using a lot
- 14:44of historical data to kind of, you know, cut the process of having the idea to completion.
- 14:51I mean, Tanya, how much savings have you found on the business development side
- 14:57of being able to go after business, complete an RFP?
- 15:01How are you applying the technology to allow you to reduce the workload on the
- 15:08business generation side of things?
- 15:10I would say that right now, backing that up, before we implement anything,
- 15:16we do look at that business case, right?
- 15:18Like, what is it that we expect this to save us in terms of time or the value
- 15:24we expect from a product so that we can measure, are we actually getting this, right?
- 15:31We don't have a system in place for a long enough period right now,
- 15:36like a full years, for where we've had full training, implementation,
- 15:40benefits to measure year over year to give it an exact number.
- 15:45But I know, for example, with the Shred AI product that OpenAsset has,
- 15:50I did a lot of calculations on that and have been monitoring from our group.
- 15:56They were in their early adaption and use of it that I expect that that team's
- 16:02going to save 30 percent in terms of time.
- 16:06And, you know, we have a large marketing team that's massive, that's people.
- 16:10Right and and allow just speaking
- 16:13strictly on the marketing size and even in business development those
- 16:17are roles that traditionally in an ac firm
- 16:20are very reactive because it's hard
- 16:23to find time to be proactive and the goal here is not only a better product
- 16:29but to allow people to capitalize on their most valuable asset which is their
- 16:34time so that we can do that more value-driven work right that thinking piece
- 16:38and reduce, you know, okay.
- 16:41Shredding an RFP, right?
- 16:43Like all of those pieces that can easily be automated.
- 16:47Yeah. And I think that's, in my experience also, using it in how we've applied
- 16:54it here at ACEC, it is freeing up that time to be able to do other things.
- 16:59It's that more strategic top-level thinking that you can do because some of, yeah, some.
- 17:06Pretty quickly out of the gate by applying it, applying AI to it.
- 17:09It's not going to replace anyone, but it's going to allow those people to have
- 17:13more time to think at a broader level.
- 17:16Eric, how, what's your take on that? I mean, how are your tools kind of designed with that in mind?
- 17:23Yeah, it's exactly designed with that same philosophy in mind.
- 17:26Like AEC marketing teams are often very small, but mighty.
- 17:31So it's more about how do we leverage AI to remove some of the tedious manual
- 17:36tasks that they are burdened with to free up their time to be a lot more strategic
- 17:41and creative, which is really what they do best, right?
- 17:44So a lot of what we do with Shred AI is being able to automate some of these
- 17:49manual tasks and also to help provide with surfacing content and information
- 17:54that would otherwise be buried under pages and pages of documents and PDFs.
- 18:01When you talk about RJC, like Tanya said, they've been around for 77 years.
- 18:05When you think Think about 77 years worth of content to dig through in like
- 18:09a SharePoint server, for example.
- 18:11It's nearly impossible to find things in a needle in a haystack type of way.
- 18:16So a lot of what we're doing with Shred AI is how do we give people access to
- 18:20that information without having to reach out to somebody who knows somebody
- 18:25that might have worked on a proposal seven years ago,
- 18:28giving you access to all that information at the touch of your fingertips via
- 18:32conversational kind of chat interface.
- 18:35That really just speaks to your entire proposal library content.
- 18:39So those are just some of the examples that I think aligns with the philosophy that Tanya mentioned,
- 18:44as well as what you mentioned around how do we enable AEC marketing teams to
- 18:49work on things that are a lot more strategic and creative as opposed to some
- 18:52of these tedious manual tasks.
- 18:55I mean, we've all been there, right? You have a question and you realize the
- 18:58person with the answer left the organization five years ago.
- 19:02The files aren't organized. You have to scramble and waste time,
- 19:07a lot of frustration and angst to try to get answers to things and to be able
- 19:12to have this repository of information that it's kind of your,
- 19:18you don't have to worry about that institutional knowledge, right?
- 19:21It's, it's, it's, it's definitely a savings for our own sanity, right?
- 19:27You know, how did you, Danny, with that, with that much data,
- 19:31I mean, some of it must have not have been digitized, right?
- 19:34You must have had to have gone through the process of getting that into a digital format.
- 19:39I mean, how did you approach just that when you talk about digital asset management,
- 19:44it's that creating that workflow to allow you to have a consistent way of putting
- 19:50in information so that you can do those natural language searches of a bridge
- 19:55at sunset and you get the appropriate things.
- 19:57I mean, how much of a culture change or was it a culture change for your organization
- 20:01to kind of get to that point?
- 20:03Yeah, I think it depends on what those assets are. So things like,
- 20:08you know, photography assets, no.
- 20:10Like they didn't even know what happened, right? It just happened.
- 20:14We just put this layer on top that now allows us to search those assets.
- 20:18But for bigger things, like more of like the project data, the fee data,
- 20:23proposal data, et cetera, that's really been a challenge that we've had to address, right?
- 20:29I feel that a lot of folks, whenever they start to look at AI,
- 20:33they try to move right into implementation and they underestimate the importance
- 20:39of data quality, like exactly what you're talking about, right?
- 20:43AI is really only as good as the information that you give it. And.
- 20:48At best, it's really risky if you don't have that data set up correctly.
- 20:52So firms like ourselves, you need to invest in clean, structured,
- 20:58and protected data if you want AI to deliver real value.
- 21:02And how that should be structured and how that should be set up and getting
- 21:07consistency across your data streams, right?
- 21:11So we're quite honestly, like, have been moving in, I'll call it blocks of where this is happening.
- 21:17But we're in the process right now of really doing that big data migration and
- 21:26that big data cleaning so that we can fully utilize AI across a variety of systems.
- 21:33And I think this is where the structure is so critically important.
- 21:39The fact that you did bring on someone who would run the data operation,
- 21:43the fact that you do have an AI working group. You know, we've done research
- 21:48with the ACEC Research Institute on AI adoption.
- 21:51We've done kind of look-ins to different ways that firms are doing things.
- 21:56And it's interesting the number of firms that don't have positions that are
- 22:04very data-centric, that are chief data officers or chief data security officers.
- 22:10And I think that's going to be one of the things that is going to define success
- 22:15for firms in the future is who has been strategic at the top level to create
- 22:20the positions to enable the collection, management,
- 22:24security of data.
- 22:28Eric, what are you seeing in the marketplace with AEC firms in that area?
- 22:33And Tanya, how important do you think that was for the firm to adopt early?
- 22:40So you can, to your point, you don't rush into adoption.
- 22:43It's that strategic consideration of how are we actually going to use the technology
- 22:47and who's going to control it?
- 22:49Yeah, I can speak to kind of like what I see out there. I think organizations
- 22:53that have been ahead of the curve in this regard are definitely more set up
- 22:59to succeed, I guess, in this future world.
- 23:02To Tanya's point, the whole concept about data and the integrity and quality
- 23:07of that data, it's the same thing of like garbage in, garbage out, right?
- 23:11So the more that there is a strategic alignment across the company that really
- 23:17prioritizes making sure that the data is clean.
- 23:20But also making sure that there's a collaborative working group like RJC has
- 23:25that allows the different functions within the company to make sure that they're
- 23:30tying and connecting the dots together to allow AI to also help map that data,
- 23:35I think is also really important.
- 23:37So as an example, when you think about, you know, leads within a marketing department
- 23:42that then gets converted into a sales quality qualified lead in the sales department,
- 23:47in and then eventually to close,
- 23:49the labeling and the way that you do your formatting all needs to somewhat align
- 23:54so that AI can connect the dots quite easily across the journey of a lead, for an example.
- 24:01But you can use that analogy to apply to pretty much anything within their organization.
- 24:05So just making sure that there's almost a data sanitization process to make
- 24:11sure that all the different functions and the data that weaves between them
- 24:14is all well connected through good data quality, I think, is obviously really important.
- 24:20So that's kind of what I see some companies doing really well.
- 24:24Yeah, same. I feel that more and more firms are doing a lot of like what we're doing, right?
- 24:31Bringing in data analysts, bringing in more business service professionals to
- 24:36lead this and to help support this.
- 24:39We even have a project manager that's dedicated to transformative projects is what we call them.
- 24:45So that's everything from how we're rolling out and preparing for a new CRM,
- 24:51for example, or an ERP system,
- 24:53having people that it's not off the side of their desk anymore,
- 24:57and that they're actual specialists, like a knowledge management coordinator,
- 25:01that people that understand their role and their function and those possibilities
- 25:05and are just sort of learning it, but can guide the firm and give that support.
- 25:11It's really fundamental when you get to a larger scale, right,
- 25:14if you want to make things scalable.
- 25:17I think that's a really good point because what we're seeing in some of our
- 25:23research about firm of the future and what is the engineering firm of 2030 going to look like,
- 25:282035 going to look like, I think we're going to see a lot more support positions
- 25:33in project management that are definitely information and database that are
- 25:38supporting the project teams.
- 25:41From beginning to end of project to coordinate, organize, corral all the data,
- 25:48maintain the structure and integrity of the data, and actually allow these AI
- 25:54programs to do what they do because they're kind of in charge of it.
- 25:58And it's going to be that interesting diversification of talent within a firm.
- 26:03You're always going to need the PEs to sign the plans.
- 26:06You're going to need somebody who's physically working on the work and making
- 26:09sure that it's done the correct way.
- 26:12But there are going to be a lot more people who are actually in that support
- 26:15role because we are not at that point, to Eric's point, of being completely
- 26:19autonomous and allowing systems to do their thing without human intervention.
- 26:23What do you guys think is probably the three best tips that you could provide
- 26:29for firms that want to get started here?
- 26:31Because you have a lot of best practices that you talked about here.
- 26:34But if you could crystallize them down to like, you know, maybe three of the
- 26:38best things you could do at the outset.
- 26:40Tanya, what do you think it would be? And then Eric, from your position at Open
- 26:43Asset, what do you think? I'm going to start with what I think is like...
- 26:49You're starting. Today is day one. You want an AI policy in place.
- 26:53Most firms, I think, you should know, like your employees are probably using
- 26:58ChatGPT, but you should make sure that they're using it responsibly, right?
- 27:03And have something to stand on and help educate people and back that up.
- 27:07I believe that would be the first thing I would say is most important is an AI use policy,
- 27:14something that provides employees and protects the firm too with what is compliance,
- 27:20what is data security, what is ethical use.
- 27:24And then pairing that like we did, I feel it's working very well with that AI
- 27:28user group that's guiding that adoption, right?
- 27:32And helping to set those measurable goals that you spoke of earlier, right?
- 27:35Like understanding, like, what are the KPIs? Like, what are the returns that we expect out of this?
- 27:42Second, I would say, is change management.
- 27:45And I feel that's one that's often overlooked. Like a lot of times we'll just
- 27:49throw in a tool and we don't take the time to invest in the user adoption, right?
- 27:57So bringing about what is this change? What is the value?
- 28:01Giving people the training and the support that they need so that it can have
- 28:05that real impact that we're looking for.
- 28:07So training people, like if it's a chat GPT even or Microsoft co-pilot or even
- 28:13open asset, like how do you prompt it correctly? so that it's really effective.
- 28:18And then also teaching people, like you do need to critically review this,
- 28:23right? And what those outputs are.
- 28:24There has to be a human element, as Eric mentioned.
- 28:28And then third, it ties back to both of those, would be making it measurable, right?
- 28:35Knowing what is the outcome that you are looking to achieve and measuring that.
- 28:41You don't just walk away, right?
- 28:43Is it doing what it says it will do? Are we getting the benefit we expected? And if not, why?
- 28:49So yeah, I guess I'd be really looking, continuing to look at the business case.
- 28:54Those are those are three good, good tips that I think anybody out there listening
- 28:58should take to heart. Eric, how about how about you?
- 29:02Yeah, I love Tanya's. I think that the training and the measurable results are
- 29:07two fantastic ones. I think on the tech product side, I think the first one is around.
- 29:14Making sure that we're designing products with really intuitive design.
- 29:18AI is obviously new technology.
- 29:20There are some obviously people that are very experienced and experts in prompt
- 29:25engineering and typing up really great prompts,
- 29:28but it's the way you prompt AI and the results that you get can vary really
- 29:33greatly depending on how well your prompt is structured, right?
- 29:36And I think it's not the best to always expect people to be prompt engineers, for example.
- 29:42So how do we make sure that we design products that are so intuitive that people
- 29:47can still get fantastic results without having to be a prompt engineer as a
- 29:52side hustle in their, you know, aside from their day job, right?
- 29:56So I think number one is knowing the technology, knowing the variability around
- 30:00how the technology performs and making sure that you're designing your products
- 30:04with really great design so that anyone can get really effective use out of
- 30:09it without having to just be an expert in AI. So I think that's the first thing I would recommend.
- 30:15I think the second thing is around creating AI that's very industry specific.
- 30:20Like for open asset, we've always been very AEC focused. That's always been part of our DNA.
- 30:26But AI off the shelf, whether you use ChatGPT or Claude, it performs really, really well.
- 30:32But when you dive into specific verticals and industries, whether it's legal,
- 30:37whether it's medical, whether it's AEC, that's the nuances of the AEC industry
- 30:42is very unique, obviously.
- 30:45So I think for us at OpenAsset, we're always trying to create kind of custom
- 30:50bespoke AI that's very, very fine-tuned and specific to the AEC industry,
- 30:55just to make sure that it's a lot more effective.
- 30:57It can detect things and recognize things that your regular off-the-shelf AI
- 31:02wouldn't necessarily be able to.
- 31:04So I think the second thing that we're spending a lot of time on,
- 31:06time and focus on. And my second recommendation is to create very specific industry-specific
- 31:13AI that will be just a lot better than a generic, you know, vanilla type of AI there.
- 31:18And I think the third thing is making sure that you're adaptable.
- 31:21I think within tech these days, especially with generative AI.
- 31:25Things are changing and moving extremely rapidly.
- 31:27There's new models that are coming out from different providers on a week-by-week, monthly basis.
- 31:32There's new open source technologies out there that's appearing every single
- 31:36day, how do we make sure that we're agile and we're fast and we're adapting
- 31:41to the constant new technology changes,
- 31:44but also adapting to the needs of the industry and moving at this breakneck
- 31:48speed to make sure that we're constantly evolving and iterating on our products and our solutions.
- 31:52I think that's the third thing that I would recommend is just making sure that
- 31:56you're constantly adapting to the changing landscape that is both tech and AAC these days.
- 32:02Yeah. Those are also three good examples.
- 32:06I always feel like these conversations can go on forever because there's so much to dig into.
- 32:12Just listening to that, thinking about doing a prompt engineering 101,
- 32:15because it's not as though you just walk into it and know how to do a prompt, right?
- 32:19And you're right. Your results vary widely by what you put in,
- 32:25garbage in, garbage out.
- 32:26So there's so much to dig into.
- 32:29We covered a lot here. I mean, is there anything else that you want to bring
- 32:31up, Tanya or Eric, that you think our audience should know from this conversation? Yeah.
- 32:37One thing that I should have mentioned earlier that really brought
- 32:41forward the case for AI and the recognition to our executive of the importance
- 32:49of this was speaking to a CEO of another engineering firm in the space who be
- 32:57viewed as a very traditional engineering firm.
- 33:00They said that they were making a strategic shift, that they're no longer an
- 33:04engineering firm, they're a data firm.
- 33:06And that's the level of importance that this is today in our industry.
- 33:12And I see that in ourselves.
- 33:15And I know that is true of many others.
- 33:18And there are such incredible opportunities beyond the ones that I think we're
- 33:23all looking at. We're all seeing like, oh, yeah, we can automate minutes now.
- 33:28We can, you know, do virtual sessions like we're doing here today and then have
- 33:31something generate the notes and the takeaways that there's other opportunities
- 33:36that going to what Derek was saying about industry specific.
- 33:40Think of like what is firm specific and the
- 33:43opportunity that we have that really sets us apart
- 33:46right that can bring that value to a
- 33:49client if you have tons of data about
- 33:53lifespan of a certain material type for
- 33:56example or about a portfolio of buildings like how can you use ai to help inform
- 34:04decision making with that right to bring next level value there's so much that
- 34:08we're not even aware of today that's going to be happening for us all tomorrow.
- 34:14And we shouldn't be scared of it. Yeah.
- 34:16Yeah, that's great. I remember sitting in a meeting at FIDIC,
- 34:21the international group for
- 34:23engineering companies, and they had a former Silicon Valley engineer who,
- 34:29in 2019, and he was saying, you're not an engineering firm, you're a data firm.
- 34:34And the response from the audience then was a mix, right?
- 34:39Because we were right at the beginning of the kind of AI revolution.
- 34:43And I think that that audience would have a definitely different reaction hearing that today.
- 34:49So that's a great, great takeaway. Eric, how about you?
- 34:54Yeah, I think similar to Tanya's point, I think right now we're still only scratching
- 34:59the surface here, right?
- 35:00So when you think about how the general public sees AI and interacts with AI
- 35:05these days, it's primarily in really just two forms.
- 35:09It's automation and things that it does on the background, or it's chatting
- 35:13with it from a conversational UI perspective.
- 35:16But the next wave of AI is really around how do we have agentic workflows where
- 35:22we have multiple AI agents that are all collaborating together to complete an
- 35:27end-to-end workflow here.
- 35:29So when you think about that world, that is trained AI agent specialists that
- 35:34are really great at doing that one task.
- 35:37So you could almost think of it as an assembly line where you have a bunch of
- 35:41AI agents that are then completing their individual tasks and then passing that
- 35:47information on to another AI agent.
- 35:49Once you start to daisy chain and connect complete flow using agents,
- 35:54that's where it really unlocks the power.
- 35:57So no longer are you having a one-to-one interaction chatting with AI.
- 36:01But you're actually having AI do a whole series of connected tasks for you to
- 36:06just really, truly automate things and really reduce manual labor.
- 36:10So an example of this is you could leverage AI to not only find content from your past proposals,
- 36:16but then pass it along to another agent that will then take that information
- 36:20and create a first draft proposal for you and then even design the proposal on your behalf as well.
- 36:26So being able to just think about workflows in a more meaningful way and how
- 36:30you apply AI, that's really the next wave that we're talking about here.
- 36:33So I think, yeah, I think there's a lot of really promising aspects about AI
- 36:38that we're only scratching the surface on these days.
- 36:40Well, if that's the future, then all the work that Tanyan and RJC is doing is going to pay off.
- 36:45So anybody listening out there who is at the beginning of the journey or trying
- 36:49to figure out exactly how to approach this, I think the big takeaway here is
- 36:54a measured strategic approach, get leadership, buy in at the beginning.
- 36:59Get a working group and get somebody in there who can manage this and then figure
- 37:03out what the business case before you just go in and jump into the deep end.
- 37:08A lot more we could go into.
- 37:10I think we should have some follow-up programming on some specific assets and
- 37:15aspects of AI implementation, but we'll leave it there for now.
- 37:20Tanya and Eric, thank you so much for being part of the program.
- 37:24Really enjoyed the conversation. Yeah, this has been great. Thanks for having
- 37:27us. Yes, indeed. Thank you for having us, Jeff.
- 37:29Fantastic. And again, this has been Engineering Influence, a podcast from the
- 37:33American Council of Engineering Companies.
- 37:35Before we go, I'll make a plug. I know that we get a lot of views on YouTube,
- 37:40and a lot of you are watching but not subscribing. So take that time now.
- 37:45Hit the notification button. Ring the bell.
- 37:48Hit the like. It helps the program immensely, and I appreciate you for watching.
- 37:53And again, this has been Engineering Influence Podcast from the American Council
- 37:57of Engineering Companies.