211: ‘Discovering the Value of AI Through Experimentation’, with James Martin, Jess Purcell, and Christopher Parsons

211: ‘Discovering the Value of AI Through Experimentation’, with James Martin, Jess Purcell, and Christopher Parsons

Jim Martin: Welcome to the TRXL Podcast. I'm Evan Troxel, and in this episode I welcome Jim Martin and Jess Purcell, joined by our KM 3.0 series, partner Chris Parsons. Jim is the Chief Information Officer at Shepley Bulfinch. Jess is a design technology leader who's been translating messy real world workflows into teachable practices.

And Chris, as you know, is the founder of Knowledge Architecture and our guide for this AI powered knowledge management series. Today we open up Shepley's Playbook, 150 years of firm history, 11 years on their Finch intranet, and talk candidly about how they're discovering the actual value of AI through disciplined experimentation.

Today we explore how to put process around unpredictable AI so results become repeatable. Why a video first approach is slashing the time it takes to document complex workflows, how AI search is unlocking institutional knowledge buried in recordings and what it really looks like to run knowledge management without a KM department.

We also get into knowledge exploration. Trust, but verify practices and why Shepley avoids the punitive traps of traditional LMS platforms while still building durable learning pathways. One of my key takeaways from this conversation was working out loud. The team at Shepley uses social enterprise behavior, not software to shape its culture.

And inside Finch, Jim and Jess show how thoughtful posts, lightweight campaigns and visible mentorship create real engagement, not performative activity. From turning hallway conversations into firm-wide mentoring templates, to using analytics to spot bright spots instead of compliance gaps, they're proving that culture changes come from participation, not policy. What struck me most is how much this grassroots transparency combined with leadership's restraint to not over systematize it, keeps curiosity alive across the firm. It's an elegant reminder that knowledge management is really about people making their thinking visible. As usual, there's an extensive amount of additional information in the show notes, so please be sure to check those out for links to connect with our guests and for more resources, including links to the other episodes in this KM 3.0 series. You can find them directly in your podcast app if you're a supporting member of TRXL+, and if you're a free member, you can find them at the website, which is TRXL.co. Lastly, you can really help the podcast by, of course, sharing these episodes with your colleagues and by commenting and sharing on my LinkedIn posts.

You can also leave a comment over on YouTube and engage with me and the other listeners there. This was a great conversation. So now without further ado, I bring you Jim Martin, Jess Purcell, and Chris Parsons.

Evan Troxel: today we're talking with Jim Martin, Jess Purcell, and of course Chris Parsons back for another episode of our KM 3.0 series. And I'm very happy to finally get these people on this podcast. It's been a long time. You've been on my wishlist, Jim and Jess for quite a while.

and this is a great subject to be talking about. the work that you have been doing for the long haul at Shepley Bulfinch, I think is really gonna be shone a light on today. So, Chris, introduce our guests.

Christopher Parsons: Yeah, I think long haul was a great, great transition. Um, Shepley Bulfinch is a firm that I've known for a long time, and they have a long track record of innovation in general, but more to our point on this podcast, innovation around knowledge management. Um, so Shepley has been a client of ours since 2014. Their intranet is called Finch. This is the first of many bird themed references coming in this episode. Just a spoiler alert. Um, Jim Martin is the Chief Information Officer, and Jess Purcell is our, is the design tech. I almost said our design technology manager. This is how, like closely you hear that I was amazing. Um, Jim and Jess together, um, as well as Shepley writ large. They've been huge contributors to the knowledge architecture community for over a decade. And as we kind of keep moving through this, welcome to KM 3.0 series where we look at AI powered knowledge management. Um, I had to bring Shepley on, um, to really kind of talk about their recent AI work, um, their knowledge management work, a lot of the stuff they're doing with video and ai, which is where we're gonna start today, as well as, and I think, I hope we get to this and can spend some time here.

I think culture at Shepley around experimentation is something I've always admired. Um, they're just. They have an idea and they run these very thoughtful experiments, and some of them they keep doing that become programs as far as Shepley Bulfinch, and some of them they retire and they're just always experimenting and trying new things.

And then they've been super thoughtful and and transparent and generous in sharing that with our community. So welcome Jim and Jess to the welcome to KM 3.0 podcast.

Jim Martin: Thanks for having us.

Evan Troxel: I I want to ask a quick question before we jump into like, what's actually going on with, KM 3.0 at Shepley and that is, okay, so one thing I know about Jim for sure is like skeptical. Cautious, right? Because there's a, the fire hose of tools and I, I can't even imagine what it's been like, Jess, to work with Jim through the AI thing. so the things that Chris was just talking about with innovation and kind of this methodical strategic, I know that is, those are the words that describe how it works at Shepley with this. And Jim, I'm just curious. Like in the, in the, the sphere of the cloud, the, the cloudy muck of ai. What's that been like for you? Have, have your views changed on this over the last few years? If so, how are you still skeptical? I assume you were skeptical at the beginning because you have been with so many tools over the years. Right? So in in, in a really good way, I mean that in the absolute best sense, so.

Jim Martin: Yeah, I think the, uh, the AI stuff has been really interesting 'cause it's. It is been immensely over promised. I mean, the hype is just ridiculous. But from the very beginning, from the very earliest tools that we had access to, it was really clear that it was able to deliver some value. So, you know, uh, as Chris mentioned, we are um, a culture of experimentation and we just started trying to figure out, well, what actually, where is the boundary between what everyone is saying that it can do and what can actually do that?

It delivers value. Um, and we're constantly looking for that boundary. And that boundary is, is all over the place. It's almost like asking someone, well, what is the value of the internet? All right. Well, it depends on a lot of things and it will change tomorrow versus today. And, um, some of the times we find value in things and the limit of the value is because we didn't imagine hard enough.

And sometimes it's because the capability didn't exist. Uh, but we also find ourselves going back and rerunning old experiments to make sure that nothing has changed because so much of the models have changed. Our understanding for what it can actually do can change and all the rest of it. Uh, that said, it really annoys me that you can't ask it the same question two times in a row and get the same answer.

Evan Troxel: Totally.

Jess Purcell: Yeah.

Jim Martin: But if the answer you get is good enough in both cases for the use case that you're experimenting with, then it can still add value.

So it's really about trying to understand what is it good for, what do you need, and, uh, finding all the ways that those Venn diagrams overlap in a meaningful way.

Christopher Parsons: I think that, um, not to be too pedantic, five minutes in, but I, I would, I think, I think skeptics are great and pessimists are the ones we wanna watch out for. Like, I think for anybody innovating, like having skeptics on your team is really great 'cause they can help counterbalance the wild-eyed optimists, you know, that aren't seeing any downside and just wanna charge ahead.

So I feel like

any great team should be, should have a blend

there. I think pessimists just bring the whole vibe down, but like a skeptic that's asking hard questions and that's cool, but where's the value?

Like, I think that's a really valuable thing to have.

Jess Purcell: Yeah, I think that's what we've found with a lot of AI tools in the a EC industry and really probably at large it can make stuff, but can it make the thing that you want it to make and need it to make? And so once you apply it to an actual project, it just all falls apart.

Jim Martin: Yeah, if you get lucky and you get the thing that you wanted or you didn't know what you wanted and you were just taking what you got, uh, and it's better than what you had then fine. But it's like using it as a directed tool in order to achieve an end, uh, is a lot more work and is a lot more variable in its ability to produce.

Evan Troxel: And because you get a good outcome, once in a while, it makes you think you're gonna come back and get a great outcome the next time. And, and that's just like you said, Jim, you're gonna get a different answer, even asking it the same question two times in a row. And so I think that's frustrating for a lot of people too, right?

It's like. Wow. It gave me this great thing first shot, and then I go back the next time and it take, oh, that didn't work. Oh, that didn't work. Oh, that didn't work. And that, and then that, that's like this, it takes a lot of time because you never know the path that you're gonna be going down when it comes to working with these tools.

You don't know where it's gonna take you, and you're not, it's not gonna be a direct line like it was the last time. It's gonna be a, a squirrely, squirrely line for sure.

Christopher Parsons: I mean, as a builder of these tools, you know, every, every release of, for example, our AI search that comes out, there's less and less AI in it, and there's more code around the AI to help make it more predictive and more deterministic. So there's this like really powerful, like almost like fire at the center of the product, right?

But you wanna surround that powerful. Fire's a decent metaphor. I mean, we had, we had a lot of problem with fire in buildings and in society in general until we got some like safety mechanisms and understood how to contain it and use it, you know, more proably. And I think this is the same thing I think every, as someone that's building it, like we don't like that either.

We don't like the slot machine, you know, uh, kind of effect. And so like, how can we continue to use ai, what it's good for, and then kind of surround it with deterministic code that helps make more predictable outcomes, you know, wherever we can.

Jim Martin: Yeah, we, we don't do as much code. We do process, but we do the same thing. And it's like

if you have an unpredictable tool, you need to build process around it in order to make it more predictable. So you design the process and use case and as much as you can in order to bring it into a narrower focus and have a more reasonable expectation that when people come to use it, they will find repeatable value.

Evan Troxel: and you work at an architecture firm, which is, I mean, how many times have you heard like, this is the way we've always done it. And so, and, and, and that process is like, okay, that process is on lock. Now we're just gonna use that process and we're gonna, you know, okay, that's autopilot. Now we can focus on, but, but to your point earlier. Things just keep changing with this. And so the, I assume the processes also need to keep changing or just getting looked at again, because what didn't used to be possible is now possible. For example, there's different ways of doing things. There's new features added. There's all of this constant churn of change that's gotta be difficult to deal with in a large organization.

Jim Martin: Yeah, it is. I mean, Jess, you should talk about our Miro board.

Jess Purcell: Oh yeah, so any experiment with ai, we do, at least with my team, we document the whole thing in a Miro board, and it is actually split up by tool and it's dated. So you can see some of these tools, like we drop it after, it's like not good. Two years later, we've picked it up again and you can see it's actually giving better results. But the nice thing about Miro is it's open to anyone in the firm to actually go in and see what people are working on. Um, what can you actually get outta it? What are our thoughts as, uh, geeks? We call our IT department geeks, uh, for context. Um, and yeah, so we're, we're very transparent too. I think our policy around AI is very much like experiment, but understand the, like, trust, but verify, you know, what tools are safe, what aren't.

And that's something we're really struggling with too, is like, how do we communicate that to the firm, especially when things are changing, like Midjourney now if you upload something, they will train the model on that, that's no longer protected.

Evan Troxel: Moving goalposts

Jess Purcell: Mm-hmm.

Evan Troxel: everywhere.

Jim Martin: EULAs.

Christopher Parsons: Jess, you, you spoke at KA Connect last month.

Jess Purcell: I sure did.

Christopher Parsons: for people that Dunno, that are listening, that's our annual knowledge management conference for AEC, Jess gave a talk called How Video Became a KM Superpower at Shepley Bulfinch.

Jess Purcell: I think that title really oversells it, but

yeah,

Christopher Parsons: I don't know, people were pretty into it.

Jess Purcell: they were, I was surprised, but I think it was, I think sometimes the simplest ideas are the ones that really resonate with people because it just feels like, oh yeah, absolutely. That's the thing that we need to do. So,

Christopher Parsons: Okay, so whether oversold or not,

tell us about this. KM tell us about video and why, why it's important, why you were able to talk about it for 40 minutes.

Like what is it, what is it about video that you, that you guys are doing?

Jess Purcell: Yeah. So I think it's two parts. We're 150 year old firm, so we have 150 years of stuff. But we have about 15 years of digital stuff, uh, 11 years of stuff on Finch. And so we have like 300 plus videos of just knowledge we've shared trainings we've done in person. Uh, and so the AI search that Chris has come up with, uh, really enables us to uncover information in those recordings that you wouldn't think of. Um, things about like random history within the firm. You would think like, oh, that's not on the intranet, but someone made a comment about it in like a podcast style, like principal interview. And so like you can ask a question about it and you get an answer. And we're finding a lot, a lot of people, um, like asking questions that like, I would not expect to see the answer and they're finding the answer, which is really interesting. So that was kind of the first half of like. Taking what we already have and making it discoverable. And then the second half of the talk was around our experiment in moving towards, uh, what I call video first content creation. So I am personally very obsessed with YouTube, um, and also a little bit Instagram, although I find it annoying these days.

Um, so the idea of like, can you distill down something you wanna do into a short video? Someone's scrolling through Finch and they need to like know how to do something, or they see something interesting, it's two minutes. They can watch it and then move on with their day. It's not like a tutorial. They have to click through, click somewhere else, watch a whole thing. So we started experimenting with that. And then we started experimenting with doing some longer form content, some video on demand stuff. Uh, and what we're doing lately, now that we have the AI search, is we actually build the video first. So you spend a short amount of time doing like a prep session.

You're like, these are the things we want to go over. You get an expert on the call, they talk through it. You can actually see screen clicks, which I love. And you can actually see everything on the screen, not just what's been screenshot in a written, uh, guide. And then you can use something like Copilot or ChatGPT, or sometimes we use like the Finch, the Synthesis AI to actually write a technical guide to go with it. And so instead of asking like the internet or like. AI basically to make you a thing from the internet. You're asking it to summarize like content that you have built yourself already. Like you've already said the things, this is how we talk about how we do things at Shepley, and then it builds the written content around it. So we've been, I did a whole thing of interviewing all the people who have been building this kind of content, and they're all the same people that usually write like all our written content, did a little like comparison, and it's like a fifth of the amount of time to do it. And then if you think about it, like your content experts don't need to edit the video. That could be someone in like a level one position, or in my case, like our design technology specialists. So those people, then it's even less of an ask on your content experts. It's like 20th of the time. It gets it down to like one to five hours of their time. And then you have the people who are editing or consuming that content 'cause they're watching and watching and watching while they're doing the edit.

And so you're building that knowledge within someone, um, who could become like the Nexpert as Chris likes to use.

Christopher Parsons: Can you give an example, Jess, of like some of the things that used to be like a hundred or 200 hour efforts for these written guides that you're

now doing video first and compressing that.

Jess Purcell: for sure. Um, one of the examples is like Autodesk Construction Cloud, like just project startup. You would have like seven or eight guides of like setting up a project. This is our folder structure. This is why things are a certain way. They are, this is how you upload a model and that can be distilled down into eight, three minute videos. And then you have the AI write a guide within like an hour and a half. And so even though like that one for example. The guess is like 200 hours to write something like that. Like that's a lot of content and it's very technical and it's a lot of screenshots. And then peer reviewing is a big part of it too. So someone having to actually like read through the thing, try it, see if you're missing anything. Whereas like the reviewing, watching a video, like you just watch it and you do it and you're like, yeah, that, that worked.

Christopher Parsons: the writing's so slow, right? Just 'cause you have to

stop. You don't, you don't have all the stuff memorized. So then you're like doing it on one screen, writing on another

screenshotting, and then for someone to QA it, they have to actually then go do all those things too. Versus if I'm watching, if you recorded it, Jess, you don't have to think about that stuff, but I'm also watching you do it.

And I'm like, well, unless she completely invented a ui. I don't need to go

Jess Purcell: Yeah.

Christopher Parsons: Right?

Evan Troxel: Right.

Jess Purcell: I think also like us as technical people, we get really into like writing the perfect way to describe how to do something. Versus is like someone recording you do it. Um, that's the other thing too, and I was talking a little bit with Susan at KA that the prep meeting and the facilitating is probably part of what makes it faster is you have someone who's kind of like, rather than asking someone like, Hey, you're the expert at detailing doors, teach this class.

It's like, oh, we wanna run a class on these five things. Can you speak to it? And I think our people don't necessarily think of themselves as experts, even though they are, but they know that they're good at specific things. So you can kind of prompt out exactly what you need them to show or say, uh, in a video really quickly.

And then you just kind of make them look good, uh, clean it up,

Evan Troxel: Jim, I have a question for you regarding this. In, in the culture of kind of continuous learning, which is always this thing, you hear those buzzwords together all the time. We have a culture of continuous learning. Our, our being an architect is just, that's all you ever do is you, you just keep learning because you can't ever know it all and things are changing.

Et

cetera.

Jim Martin: Yeah.

Evan Troxel: But you also need teachers, right? You need people creating the stuff that, that people are learning inside of the company, and traditionally that has come from the outside. So whether it's, you know, you're investing in some. Course LMS from some third party co company, or it's YouTube, right, to just as example, right?

A lot of people are just learning ad hoc stuff from people who have been willing to put it on YouTube, and that may or may not align with the way the standards of the firm. Oftentimes it, oftentimes it doesn't, right? And it's like, oh, now we actually have to train people how to do this, the, the way we do it, not necessarily the eight different ways that you could do it or the 53 different ways you could do it. So. How is this kind of, it sounds like it's lowered the barrier to entry to actually create the content and make it a lot less kind of overwhelming for, for people to share their expertise. I'm just curious from a cultural standpoint, what have you seen happen in the company with the success of this

rollout?

Jim Martin: Yeah, it's, I think, uh, Shepley's preference has been for people to go out and acquire knowledge, um, understand it through the lens of Shepley and then regurgitate it to the rest of the folks at the firm, and that, that has pluses and minuses, as does any strategy that anyone would run, uh, for something like this or anything else. The, the pluses is that we are able to be pretty consistent in making sure that the information that we are providing, uh, for training the rest of our firm is not inappropriate, or not sized properly, or not shaped properly for the way that Shepley wants to work. The downside of that is, is that it is an incredible amount of effort for people who are already on projects,

which is why, you know, Jess's kind of sussing out of the, our ability to use video to reduce the amount of effort on our SMEs.

Subject matter experts is so critical because. Then they can spend more time learning and acquiring all this knowledge from wherever, venues, conferences, something they've learned on the project conversation with random, uh, partners or contractors, uh, wherever it is. You know, when they find really useful information, they have a mechanism in order to convey that to the rest of the firm. Uh, interestingly Chris mentioned that we are, we at Shepley have been, um, a, a contributing member of the knowledge architecture community. We're basically doing the same thing. Uh, when we learn something that is, that we think is gonna be helpful, we will kind of share it and working out loud and exploring out loud are things that we, we do pretty much as a matter of course, in the culture of the firm.

Um, which I think is great. You. Try to find ways to minimize the downsides of any strategy you're running. And that's why the, the AI search and the video stuff is so interesting for us. 'cause it allows us to reduce the friction in the overhead aggravation for people that are, uh, we are relying upon in order to do the actual architecture, but also to help everyone else, uh, get better at doing architecture so that they can keep going out, searching and acquiring new useful information and knowledge from the world about the things that we do. 'cause Evan, as you suggested, it's like you don't get to stop learning as an architect because then you would only get to play with the tools you already know, or the materials you already know, or the processes that you already know and so much is changing so quickly that we really do not have the ability to stop learning and still be effective in our chosen craft.

Christopher Parsons: I think you said something really interesting there, Jim, that I may wanna ask both you and Jess about. So, the rate of change does seem to be continuing to increase and so, and I think you're right, you're not gonna stop learning, but you could stop documenting and you could stop sharing. And so it feels like just the strategy you described about being video first and having your geeks, you know, your DT team help support subject matter experts in reducing the lift on their end feels like it makes those knowledge assets more sustainable and durable because it's not a 200 hour effort to replace. This long written documentation, it's maybe for the subject matter matter, you said one to five hours. If ACC totally changes or your approach to site observation reports totally changes. Is that, is that part of like your conversation with them?

Jess Purcell: Yeah, I, I think a important part of context, it's not like we are building a massive repository of information, so we don't build content that exists in the world. I'm not gonna build a series of how to use Bluebeam or an an intro to Revit. What we tend to build content around is when we're rolling out processes or we identify places that people need help. And so like we did a series on field reports, but it's really around how to use like our iPads, our templates. This is how we do field reports. It's not necessarily like an intro to Bluebeam on the iPad or how to do a field report in general. Uh, same thing with ACC. That series is being built because we are rolling out a new folder template. So it's really around, this is how you set up your project to be successful within Shepley. This is our folder structure. This is how you apply permissions, that kind of thing.

Jim Martin: to your point there is there is. Part of people's reluctance to deal with change is the amount of aggravation that they know they're gonna have to go through in order to get through the cycle of change. And making it easier to create documentation for a process means that people will have less perceived opportunity loss, um, or sunk cost in, in the recreating of that content if the process needs to change in some meaningful way. So the more we can reduce the friction about creating like the, well, this is the communication plan for how this change needs to, uh, take place in the business, the easier it's gonna be for us to convince people, okay, well a new thing came out that is gonna have substantial amount of value for us and we need to change that thing that we just put into place last year. If it took them 120 hours to write that documentation for it, the people that did that are gonna be real sad. Uh, and

Christopher Parsons: gonna have a lot of sunk cost in that, yeah,

Jim Martin: they will be slightly less sad if it only took them two hours.

Christopher Parsons: Mm-hmm.

Jess Purcell: I think also building out that content into bite-sized portions, so if something changes, you don't necessarily have to throw out the whole series. You maybe just like remove a video, add a new one, or if there's like a new feature, you can just like slip that one in, in the middle. And so there's, it makes it easier to update content rather than rebuild content.

Christopher Parsons: To make it because it's more modular.

Jim Martin: Yeah. But that said, video is harder to edit than text is.

Evan Troxel: For sure.

Jess Purcell: Yes.

Jim Martin: you create, you'd create strategies around managing video, which will help it get closer to managing the, the effort to manage, uh, text and screen caps.

Christopher Parsons: It's easier to create video, but it's harder to edit, whereas the text, it's harder to create, but it's easier to edit

unless

you have to

Jim Martin: clear,

yeah, and we're doing both because, I was having a conversation with our, uh, CEO and her comment was, yeah, videos aren't as helpful for me if I know what I'm talking about. 'cause I'm really just looking to skim the document to find the piece in the text and the one screenshot that I need to bring me to the place they need to go.

Um, so different people have different preferences based on either their expertise in the particular process or, uh, their predilection for wa wading through a sequentially presented video versus a zip to the place that you need and then kind of beat on it until it works sort of deal. So we're doing both.

So there's, you have to try and figure out how to make the upkeep of that, uh, knowledge asset as easy as possible considering that you've got so much stuff baked into it.

Evan Troxel: Can I want to ask about this? You know, you have 150 years, like you said, you have 15 years of digital, but you, you have a huge storied history and a lot of information that's captured in various ways, drawings, and, you know, whether it was a scans of paper versus digitally created from the beginning, but. Something that's come up previously in these kinds of episodes is, is just the idea of expiration dates on information. How do you handle that? I mean, it sounds like there's just this constant monitoring, this is like copy monitoring in Revit. How, what, what is going on with information and how do you flag it when you, how does it get flagged when it's, oh, this needs to be reviewed?

Or is it just on a running schedule, or how are you dealing with that? Because I think a lot of people are just overwhelmed with the idea of, well, there's always just more information that we're. Teaching people about, like these things just stack and stack and stack forever. Is that, how do you, how are you dealing with that at Shepley?

Jim Martin: I would say that it isn't a closet, and the closet door is mostly closed and we ignore it most of the time.

Evan Troxel: I mean, that's super honest, right? That's

Jim Martin: When, so AI search has been really interesting 'cause it, it has provided so much more visibility into content that just kind of hangs out.

And in some cases that content hanging out is great because it is the only information about that particular subject that exists on our internet. But in some cases it's just hanging out because it is no longer accurate and hasn't been for two years. Our intranet is 11 years old. Some of the things that we put into it 11 years ago were accurate and correct at the time, but are no longer the case.

Evan Troxel: Right.

Jim Martin: So I would say at the moment our strategy is, uh, we have articulated to our folks, so if you see something weird, uh, or old, uh. Trust but verify, which is really good advice in the world of ai. Um, and bring it up. If you see something that shouldn't be there and we will address it at that time. Uh, I am hopeful that at some point we can develop more sophisticated tools to help do that proactively because it would be better if no one was ever presented from with information that was out of date or inaccurate, uh, from our internet. Um, but as much as we use it as a tool for communications, it just doesn't seem like that's doable right now. So we know that that's a problem and we, uh, struggle with that as much as anyone else does. and at some point we're gonna figure out a way to solve that problem. And we do, we're gonna let everyone else know in case it's helpful for them too.

Christopher Parsons: Some of the things in our community, people do, they do a little, and, and Shep correct me if I'm wrong, if you guys are doing this, Jim, but like they'll just set up a rotation schedule to go through the different sections of their intranet and make sure they hit

everything twice a year. You know, kind of like a, that's, that's a way to do it. Um, there are some technologies that we could put in play. Um, I think that one of the things that's so hard about it is it's, it's not always clear when something's out of date to, you know, it's not, it's not how old it is. 'cause there's things that were written 11 years ago that are still as true as they were,

you know, and something six months old or three months old could also be out of date. So it's an interesting like, um, awareness. So like as many tools and processes you put in place, some of it's just like culture and awareness and habit of like, knowing that like, if I'm making, if I'm making a change, we need to make sure that it's updated in the knowledge base. And that gets tricky to roll through an entire culture. Or the person that added the thing is no longer at the firm. And so it's a, it's, I I think that Jim's point around AI search has just helped because you can start seeing the people that, that the, some things are inaccurate and don't really matter 'cause no one's gonna see it anyway. But then it will, AI search has had this way of kind of reactively, I think the kind of more review process would be proactive, the reactive way of being like, no, this is stuff people are seeing and they're coming across this on a regular basis.

And it's out of di It helps you prioritize the kind of more reactive knowledge management side in addition to the proactive strategy, if that makes sense.

Jim Martin: Yeah, but we have, we have videos of town halls that we've run or quarterly, uh, meetings that we ran in 20 18, 20 17, and some of the stuff that's in there is great. But a couple of things we said in there is not our current practice.

Christopher Parsons: Yeah.

Jim Martin: But going into an hour and a half long video from eight years ago and chopping out the piece that is no longer relevant is a level of effort that we have not managed to, uh, build up the interest in solving.

Christopher Parsons: and on

top of that, Jim, like, and this is something that's come up at Shepley, but other places too is even if it's inaccurate, sometimes people wanna preserve it as an artifact of culture.

Jim Martin: right

Christopher Parsons: So it's like

sometimes inaccurate is fine because this is who we were back in 2014 when we were doing this

thing.

And we don't want that to change. But we also don't want people to get the wrong information. So there's all this like stuff in AI around weighting and kind of biasing towards more recent cons. You know, there's like a bunch of different strategies to take, but some of that is like, what was the thing you said in your presentation, Jess?

It's not hoarding if it's organized.

Um, and so some of it is like if we're gonna keep more stuff, rather than being ruthless and just deleting everything that's out of date, like how do we. Trust but verify, I guess is part of that, Jim, it's

like just take everything with it. You have to, people know, need to know to like question what they're seeing.

Jess Purcell: the weighting is an interesting thing because that's a strategy we move towards more than removing stuff. If something's really bad, we're gonna remove it, but it's also possible to just bury it down. So, um, for example, we had problems with file path links. This is gonna get really crunchy and technical, but, uh, all

Evan Troxel: you mean interesting,

Jess Purcell: I know, right?

All our documentation about file path links refers to new format. So if you ask a question about it. Then like you're like, oh, we're not using Newforma, therefore this is not a problem anymore and that's not true. So what? We built an explicit page with explicit language that we knew the AI would like to pick up. And so now it doesn't refer to Newforma 'cause that newer page doesn't refer to it at all and has more context outside of it.

Jim Martin: So we're, we're finding ways to mitigate the problem. Um, but I think for me, the most important strategy we are running is trying to build discerning individuals in the firm because discernment, I think is going to be a really crucial skill moving forward in the world of ai. And if we can help people understand how to deal with the, the cognitive dissidence of this is a true fact, but check to make sure that that true fact is still true, uh, feels like that's the sort of thing we're gonna have to get better at as a species in order to deal with the onslaught of information that's coming at us.

Christopher Parsons: Beyond the internet.

Jim Martin: Yes.

Evan Troxel: most current true fact, I I

I, I kind of think of a, of like a bug bounty for, for large these large.

You know,

companies that are building software, it's like incentivizing your staff with positive feedback to report these issues and let them know that, you know, raising the flag on, on these things is super helpful so that we can make sure that it, but that's, that's a cultural thing too. A lot of people are like, oh, not my problem. You know, somebody else will report it, somebody else will see it, somebody else will do that. To actually helping people own that, I think has gotta be. A cultural level kind of a thing so that the company can actually identify what those are and get rid of 'em or, or like you say, push 'em down so that they don't surface as often when, when people are doing these searches.

Christopher Parsons: actually, that's a great segue, Evan. Um, uh, so we have brought people in this program who have firms that have formalized knowledge management programs and people that are kind of halfway in between and people where it's completely informal. At Shepley, there is no knowledge management department.

There is nobody with knowledge manager in their job title, and yet you guys continue to do really cool stuff. So I'm curious if you can kind of talk about the structure of what you would consider knowledge management. You know, and so part of what everyone's getting at is knowledge management's everyone's job, which is, I agree with on one level.

On the other level it's like, but it needs orchestration and leadership and priorities and strategy and those kind of things. So I wonder if you, two of you could talk about KM at shepley.

Jim Martin: Yeah, it's interesting. I, I think you don't get to be a firm of 150 years without transferring some knowledge here and there. And you know, back in the day, Sheley, Bulfinch was able to transfer its knowledge of over, I think we were on our 10th CEO, maybe ninth. Um, and all that stuff has been happening and it's built into the culture.

And if it wasn't, the firm would not be able to survive. And until relatively recently in the firm's history, it had nothing to do with computers. Just didn't, the firm is real old, so I think we. We're pretty fortunate in that we were able to kind of come in and look around and say, okay, well we can actually use some of this cultural heritage around knowledge sharing and put a spin on it using technology in order to help it, uh, kind of happen a little bit more at scale so that it's not so much of a master journeyman sort of situation, but we can have master and many journeymen. And we've been kind of doing those experiments ever since. But I think we, we were very fortunate in that the firm itself recognizes the value of this information sharing, even if they'd never called it knowledge management. Uh, when I first saw the periodic table of knowledge management that Chris and, uh, KA put out, it was amazing because then it's like, okay, well, oh yeah, I, hmm, that's a great framework 'cause I'm already doing some of those things and I can now see in some of the things that we tried didn't work.

It's like, oh, maybe I should have added some of that in there. So I was missing a crucial element sort of thing. So knowing that it is an actual thing and that there are professional people who are super good at it, is really helpful. And I try and listen to those people anytime they will talk to me because there's always things that we can learn and adapt, um, uh, in order to make it better and easier. Um, but. Primarily the people that have the knowledge that needs to be shared. That is the most complicated to share, are in the practice of interior design and architecture. Uh, the projects are fluid. The teams are fluid. You don't necessarily have the same person you're working with forever, um, who might take, uh, interest in you and kind of take you under their wing and pour all of the information that's in their head, into your head, uh, hopefully slowly as opposed to quickly, uh, so everyone's brain doesn't explode. Um, but you know, these sorts of things are difficult problems to solve and. We need to try and find ways for allowing for everybody in the firm to learn these little things, even if they didn't have that, uh, that mentor journeyman specific relationship between two people. Because that model doesn't work in the way that we currently work, uh, in the modern day where it's much more, we formed teams as appropriate, uh, to solve the problems of our clients, not necessarily solving the problems of our people. But if we can't solve the problems of our people, we won't have people to solve the problems of our clients. And if anyone followed through that, congratulations, almost got away from me.

Christopher Parsons: And just maybe the same thing kind of on, on learning and dev. I mean, 'cause you're, how did you get roped into the, the knowledge management rodeo at Shepley? You know, like, because again, not an official title, but it's something you spend a lot of time doing.

Jess Purcell: Yeah, I mean, design technology is teaching people, right? No one cares what you made or figured out how to do. You're not adding value to the firm playing around with Dynamo on the side unless you're actually teaching that out to people. So it really is my job to teach people and in experimenting with ways to teach people, we can then codify that into a process that we can hand off to the practice to be like, Hey, this is an effective way to communicate what you wanna try to communic.

Jim Martin: well, I might just interject that, um, we practice knowledge management in our department in incessantly.

Christopher Parsons: Mm-hmm.

Jim Martin: Uh, just that might be worth talking about a little bit.

Jess Purcell: Yeah. Uh, we used to be, well, we are still in OneNote. That OneNote actually chokes. If you try to open it on a new device, you have to let it spin for like 20 minutes. But I've been pushing us to move our stuff to video and written on Finch because I am obsessed with the AI search. And the nice thing about that is the AI search will pull like our public facing pages as well as our private pages, so I don't have to have duplicate content places like this is our help desk internal content.

And this is like our external content. I can actually have that synthesize answers back to me, uh, at within one answer. Um, it's also helpful for me because we are a national firm, so our team actually works different, uh, hours for coverage. So sometimes I'm the only one online being in Phoenix for the last three hours of the day and something will come up that is like very it and I haven't done it in years.

And so like having that video to like watch someone poke through and do it, uh, is especially helpful because there are times where I don't have permissions to do stuff. So I can read a guide and then it says to do these things and I'm like, push this button. But then if I watch a video I'm like, oh, you have a different button that I don't have. yeah, absolutely a whole thing. And I guess one of the things that's important for that for us is. The way we run our design technology specialist program. We hire architects and interior designers straight outta school. We bring 'em in it for two years. We dump a bunch of knowledge in their head and then we roll them on a practice,

which is really great for delivering highly capable people into practice. But that also means a third of our department doesn't know what they're doing

most of

the time.

Christopher Parsons: You're, you're constantly upskilling. Yeah.

Jess Purcell: But that also means, I, I mean we've been doing that for Jim. How long have we been doing it? 'cause you've, you've done it even before you've came to Shepley.

Um, we've got 15 years of a part of their offboarding process is documenting everything they know. So we've got 15 years of design technology specialists leaving notes to future design technology specialists.

Evan Troxel: That's cool. That's very cool. So do you feel like you're missing out by not having a dedicated knowledge manager, overhead staff, like wrangling all this for the company that you guys are basically, this is an add-on to, to what you're doing on the technology side?

Jim Martin: sometimes. There are specific times that I can think of. Uh, during COVID was a really good example of that.

Everyone was working home from home. Um. And our IT infrastructure was going through a massive transformation and it was all that anyone in it could do, but to kinda keep the wheels on as we were moving everything from office-based to cloud-based and cloud virtual desktops and all the rest of that fun stuff.

Um, that was a massive undertaking

and we were just too busy doing any of that stuff to be thinking about any of the, uh, the KM stuff. And there were times where I would come up for air and I would be like, uh, it would be amazing if we could do some of this, this, or this. And I was able to plant some seeds and other people absolutely did some things that qualify as knowledge management that didn't require us at all. But I saw a lot of opportunities that we didn't have the bandwidth or the focus to take advantage of. And I've felt that a couple of times over the last couple of years. Um, the firm is small enough that we have. An advantage in that it's not as difficult to talk to 200 people as it is, talk to six. Um, so I think at some level we can still be effective at managing our knowledge without having someone dedicated to it. But there are opportunities that we have lost and things that we could have done, I think better if we'd had someone who is, uh, competent and focused on it rather than just doing the experimentation route. Again, there's pluses, minuses at everything and having a bunch of amateurs experimenting is got some upside, but it doesn't have all upside.

So yeah,

Christopher Parsons: I am curious about the, you've had impact, you talked a lot about the kind of KM inside DT and it, but you guys are also, there's, there's a practice, there's also HR and marketing, you know, anything from

onboarding to, you know, what are our procedures for doing things? How, I mean, I know some of the answers, but I don't know all the answers. You know, like Jim Chambers is like, I would consider part of your KM team. Right.

Your director. It's a director of practice is the title.

Correct. Um, can you talk to like how you have done knowledge management across the whole firm and into other, including some of the practice parts that you say are really hard, hard to get to.

Jim Martin: again, I think it would be disingenuous to leave everyone through the impression that all knowledge management happens in IT. Um, we can talk to you about the things that we do and the experiments that we run. And, uh, I am certainly not good at staying inside of my lane. Uh, so I will. Uh, get out and dabble.

Uh, we have a lot of experience with change management as well in it because software changes all the time. Sometimes it's our choice, sometimes it's not. So we're really good at a lot of this stuff. Um, and we share that knowledge and expertise with other folks. I know that practice, Jim Chambers and and his crew have built an entire process around, um, beta testing, new technology and new processes.

Um, for practice, we, we did a big lean kick a couple of years ago, and the nucleus of that is still going strong even if we're not speaking as much Japanese as we were at the time. Um, our p and d folks are doing a lot of this stuff, uh, all the time as well.

Christopher Parsons: P and D is People and Development.

Jim Martin: People and Development, um, are. We have a lot of people who are doing a lot of different things in the company.

It's just, but that's also a part of like, if there was a dedicated, uh, knowledge management team, I think there would be less willingness for a whole bunch of people to try a whole bunch of stuff

and kind of share what did and did not work. So if it's nobody's job, it's not necessarily nobody's job.

Sometimes it could be everybody's job and everyone could kind of, Hey, I did this, this is pretty interesting, this actually happened. Um, that sort of thing. So there are opportunities in the distributed, uh, model as well. It's not fantastic. It's not the best thing to slice bread, but it does allow us opportunities to spread around the work to a whole bunch of different people. We all are maybe doing it a little less expertly than we might prefer.

Christopher Parsons: It is distributed, but you're still doing centralized. Like Jess is still teaching people the video part, and your

geeks are still editing videos and you're teaching them, you know, so there is some like centralized support for that distributed,

you

Jess Purcell: Yeah.

Jim Martin: if there's technology involved, people will come to the geeks to get help with that stuff. So we have an opportunity to want to influence.

Jess Purcell: yeah, we also have a close relationship with Jim Chambers and his team. So Kate Rizo, who I've worked with for a while, this video process started with me, but she really grabbed it and ran with it. So like the field reports thing, the ACC, anything we're rolling out, either from IT or even from practice, is being done on this platform as we're seeing how well it works.

And it's getting to the point. And a lot of like things we do within Shepley are like this. It's very grassroots. We are starting to see groups. Who wanna push out content and release something who are starting to do video without us prompting them. So that to me is success when someone experiences something enough that they're like, this is good.

This is how we do it, and that's how I'm going to do it. I don't like running a process for more than a few years. I get bored easily. So to me it's like, if, if something was successful, it's because some, like multiple people within the firm have picked it up and run with it. And that's just how we work.

Now,

if I have to keep pushing it for five years, it's not, that's not successful.

Evan Troxel: I'm curious about the consumption side of this and, and just how it works in your firm. I know we're gonna segue into kind of like learning management and that to me kind of. Just implies that there's, there's a schedule to some of this, or at least goals maybe towards consuming certain things within a period of your career or within some timeframe, whatever. But the things that you're talking about, I would imagine are somewhat consumed just as needed, like ad hoc, like what's the latest, and so. I'm just curious what you're seeing trend-wise in your firm when it comes to consumption, because creating the content and then, and then actually having people learn from it is the whole goal,

right?

It's like you have to do one before the other can happen, but also like the consuming may not happen, right? It may not happen the way that you thought it was gonna happen, and so I'm just curious when it comes to that, what, what's the feedback you're hearing in the firm? Regarding the type of content that's being put out, and then how are they consuming it?

Are they consuming it ad hoc? Are they, are they, are they looking for learning opportunities when they have downtime versus not? I don't know. Give us some ideas.

Jess Purcell: I think it's, it's all of that. You, you're gonna have some people when you release it who are like, oh, that's interesting. I'll go watch it. You are gonna have people who just wanna watch it when they need it.

And that's part of why we're really looking towards that. Like smaller, more easy to consume kind of content. I'm gonna go do. A site visit and I haven't done CA in five years. like how do I use this template that someone gave me? Um, I think also, like you said, there are times where people aren't as busy and I think we see a lot of emerging professionals are craving knowledge. Uh, so that's an opportunity for them to kind of like dig through and see like, what do we have, what can I know about?

All of our market groups have like a one-on-one session. So if you're like a emerging professional and you're like, I think I'm kind of interested in healthcare architecture, like there are a couple videos that you can sit and watch that really kind of give you the download of like what it is you have to think about and do within, uh, healthcare.

Jim Martin: One thing that I'll maybe share that, that I thought was pretty interesting is if we have people who are on the bench who are, uh, waiting for the next assignment to start because timing didn't work out or whatever the case may be, uh, we'll oftentimes ask them to look at something in the Bird Feeder catalog, which is what we call our current L&D learning development, uh, repository. It's not really a platform, it's just a place where we put all this stuff. Um, but if you ask them to go look at the stuff, it's, it's helpful. But, uh, I have found that it is more helpful if you ask them to look at the stuff with some other meta intention involved. So it's not just go there, watch the thing.

It's like, go there, watch the thing, tell us four things that we could improve on that video.

Christopher Parsons: Mm-hmm.

Jim Martin: Um, and that helps them kind of maintain a level of focus I think, that they otherwise might struggle to have,

because they're. Looking around thinking, all right, well, I don't have my project to work on. I'm getting pestered with questions from my last project and these other things that I have going on that are my kind of passion projects for the firm.

But if you're gonna focus on this other thing, it's like, don't just do that. It's not just about your personal, uh, knowledge acquisition, but it's also about helping someone in the future by getting better information, um, in front of everything.

Christopher Parsons: Yeah, there's something Jim, in what you're saying around that, you know, uh, for full disclosure to the audience, Shepley is in our private beta for synthesis, LMS. Um, so one of the things that has been true throughout that process is that it's been very little about the technology. It's been more around kind of philosophy, culture, norms, you know, like how, how are we gonna do this?

And to like overly simplify the universe, there's kind of like knowledge sharing on one hand. And like maybe knowledge transfer on the other hand. So the knowledge sharing is, here's a lunch and learn or design tip, or go watch the healthcare 1 0 1 if you're interested. It's kind of like you can do it, like the world doesn't fall apart if you don't. And then on the other hand, there's like, you're gonna go out and do this field observation report, like we need you to know our process, or you're gonna be a project manager. You need to understand these things about billing or these things about difficult conversations with clients, whatever the things are.

And I think one thing that, not just at Shepley, but I I think you can speak to it a little bit, is I think you guys said the thing in our last meeting, which was, you know, knowledge management is for the people, by the people. You know, at Shepley we don't, we don't tell people they can't share anything.

You know, there's not like a lot of, uh, gar uh, guardrails there. But when it gets to knowledge transfer, there's perhaps a different, a different kind of set of rules like this. Actually, no, we don't wanna have it be like the wild west when it comes to teaching 'em how to do this thing. And I'm wondering just kind of your two thoughts on. On that kind of like what it means to start doing assignments, you know, or start having it be a little bit more, um, directed versus just, you know, we're a learning organization that, and that people should consume what they're interested in.

Jim Martin: Yeah, we've, we've looked at learning management systems in the past, um, and when we see one that doesn't match our culture, we kind of immediately back, slowly away. Uh, because I'm not interested in the learning management platform, changing the culture of the firm, that is not the reason to change the culture of the firm so that you can adopt a learning management platform. Um,

so yeah, so we're in an interesting situation because You know, we went on, we went on to, uh, synthesis the internet platform, you know, 11 years ago at this point. And we had the benefit of a bunch of other people getting onto the platform before us, and knowledge architecture helping us understand what the platform could and should look like. Um, so we knew what good looked like and we were able to understand how we would take that and turn it into something that would be a, a force of positive change, uh, for the firm. This time around, we are in the private beta and I think we are all a little bit trying to figure out together, you know, what the learning management system actually is going to be. So we don't have the added benefit of knowing what it looks like when it's finished. So it's much more difficult for us to figure out, well, how are we gonna actually use this tool in order to be a positive force for change, uh, in the firm. Um, fortunately it's, it's looking like it's gonna be absolutely that, but we kind of have to wait for it to be finished before we're gonna know, uh. How we would go about integrating it with the rest of the firm.

Christopher Parsons: I mean, I think, you know, LMS systems have been around for 30 years. But so many of the, the conventions that are built in come from like a high school or point of view of very punitive, or you're gonna get X grades or you're turning your homework this time.

Like it's very, I don't know, it's just like the vibes aren't great

and I think that's what Jim you're getting at. And so we're like, how do we describe, how do we design a system for high trust organizations that are full of professionals who very much are motivated to wanna learn, but then also sometimes they have to learn something. But how do we just. You just, you don't wanna turn that dial up too high where

it feels like it's, you're just getting your hand slapped all the time. And so we've been erring on the side of not doing that kind of, but I, but I agree with you, I think, we can do all the design conversations with our clients as we want.

We go build the software, but until people start using it, like that's exactly what you're describing is we're in that now, and it's like, so we had a hypothesis like, how do we not do that? But I think it's, that's on the tech side, but then I think you guys maybe had hypotheses around like, well, what does it mean to have a learning path with assignments and somebody has to complete it over a certain time?

Like, what happens when they don't,

you know, like, what does that mean?

Jim Martin: That's an excellent question, Chris.

Christopher Parsons: Right,

exactly.

riverside_evan_troxel_raw-synced-video-cfr_trxl_0297: leave it. He's gonna leave it right there.

Christopher Parsons: No, no, no. I, I mean, Ilka has the, the poet has this thing about living the questions, and I feel

like that's the era we're in right now is we're just

trying to accumulate as many of the right questions to ask as we can. Without having to try and answer them all on day one. 'cause they're un, they're unanswerable if you wanna do it right.

Evan Troxel: and I'll just want to throw in that like the, the whole, like the communication side of this equation is huge and, and. Traditionally the IT department is deploy and then put out fires, right? Like, this is how we've lived for a long time, and it's not proactive communication around the intentions and like what are the, the values of why we're doing this.

And so I'm just curious from your point of view, I mean that seems to me to be an integral part to effective rollouts, especially with things like this that are a little bit like people are wondering what are they doing? Why are they tracking? I mean there's, there's other parallels, especially in the technology side of things with analytics and tracking what people are doing in the software for good and various reasons, right?

So I'm just, I'm curious from your point where, where do you guys stand on the communication side? How are you handling that?

Jim Martin: Uh, we have a pretty robust process that Jess has been filling out for the last couple of years for. Rolling out technology change over time, and it includes multiple phases, alpha, beta, and then just early adopters and getting them to tell the stories so we're not telling the stories

and blah, blah, blah, blah.

blah. So we have a pretty robust process for that. I think for me, the bigger issue is it's like, what is this technology going to do for the processes of the firm and for the culture of the firm? And you know, that's where it's like I have zero interest in gamification of, uh, learning management. That feels gross.

And that would feel

Christopher Parsons: mean like leaderboards and that kind of thing and

Jim Martin: Yes. Yes. Um, so there are certain things that if they were options in our tools, we would simply turn them off or not use the tool. Um, I do not have any interest in making people think that because they didn't watch 57% of this video, that it doesn't count.

Um. Again, it's like there are, there are things that we could do that would be really helpful from a, a statistics point of view that would be really detrimental to the culture, So I don't want any of the tools that we build to detract from the culture of the firm that we have spent so much time building and curating because it's just way harder to build positive culture.

Evan Troxel: This makes me like wanna just point out how I think it sounds so important to have the continuity of someone like Jim Martin in a position that you're in as a taste maker who understands the culture and understands the intentions of the firm and, and what those things are. Because I think a lot of times people come into these positions and they just don't have that baseline.

To keep calling back to, to fully understand that. And so then it's, it's way less, uh, well, it's more difficult, I think, to, to make things stick because they're trying a lot more things because they're not having the taste maker say, this absolutely doesn't fit the culture. So I, I mean that, that, I just thought it might be worth pointing that out because, this is something that, that ultimately is a deciding factor for so many things that you guys are deciding to pursue or not.

Jim Martin: it is nice kind of having the history and being trusted in the, in the organization in order to do that stuff. But, you know, it doesn't require 12 years of time at a firm in order to do it. It just requires a good process.

And to this day, I have opinions about what I think is going to be good for the culture of the firm, but I also test them against reality wherever possible, so I don't take it for granted.

Um, and by not taking it for granted, I think I build up even more goodwill, uh, with the folks in the firm. That the things that I'm doing are intended to be in the best interest of the firm. And so when I make a bad call, you know, they give me the good grace of allowing a redo and I get a mulligan.

Christopher Parsons: It is interesting. We ended up here, um, we interviewed Kate Grimes from Snohetta. She will have been on, I'll try and go into the future. She will have been on last month's episode. Uh, um, and we ended up in this area of like being very, uh, thoughtful around AI adoption and understanding what certain tools will do to the culture and trying to find tools that support the culture and don't. unintentionally change it,

you know? And it sounds like we're here again now with you talking about learning and being careful around adopting tools and understanding the cultural impacts of certain technology, that they're not neutral. Um, and I'm curious, Jess, it sounds like you're the guardian of the change management process now.

Like, is that, is that kind of built into that process or is the process more like, here are the steps of rolling something out and the big picture stuff is done outside of it? Like how does that, how does that fit in?

Jess Purcell: I think it absolutely is. I mean part of it is that that's at the evaluation phase. Before we even think about testing something or rolling it out, there's always the question of like, not just is it gonna add value, but does it work within how we work and are there other tools that do that better? I think also, Evan, you were talking about the tracking thing.

For us, we build a lot of goodwill and can track a lot of stuff because of how we act on it.

Christopher Parsons: Mm-hmm.

Jess Purcell: am not interested in any tool that tracks things to the level of like what buttons are pushing, like how people are working in the tool. What I am interested in is, you know, people who are using a tool a ton that we just rolled out. Uh, that we were not expecting that person to be in it. And so it's like, oh, that's a person I can get to go post or do a show and tell about the value they're getting out of it because I am not necessarily pushing them to use it. And so they're clearly getting value out of it 'cause they're using it a lot. And that kind of message always comes better from your peers

Christopher Parsons: Mm-hmm.

Jess Purcell: than from it telling you, this is gonna help your workflow.

Christopher Parsons: So you're using analytics to find bright spots, essentially, and then

you want to amplify those bright spots.

Jess Purcell: Mm-hmm.

Christopher Parsons: Positive deviance. Is that what you said, Jim?

Jess Purcell: Yeah.

Christopher Parsons: Yeah. Right. I love that. I remember Jim, you and I, back in the KM 2.0 days of knowledge sharing coming in,

Jim Martin: So long ago.

Christopher Parsons: So long ago.

Um, you were working on a dashboard and we kinda like went back and forth on it showing like Principal contributions, I think, to the firm in

terms of their sharing. And there was like a very red version which kind of showed who wasn't sharing. And then there was a green version, which was like, here's who's sharing the most. And I'm like, Ooh, the green version seems nicer.

Jim Martin: Yeah, we should definitely do that one. Yeah. But no one wants to be below the CFO in sharing. Yeah, so when we first launched our internet, we were, we were concerned because we wanted to make sure that leadership was showing up on the internet at the same level of frequency that the rest of the firm was, so that it wasn't just seen as a place where they weren't sort of thing. And so during all of our Principal's meeting, I just, we started every meeting with a slide that showed, uh, stats for how many likes, shares, and, uh, posts that all of our leadership was doing at the time. And we talked about it the first time, and then we just left it up for a good six months, and then we didn't have to ever talk about it again.

Christopher Parsons: Hmm.

Jim Martin: uh, it was, it was pretty effective. But, you know, uh, Jess put it together for us in Power BI, and the first one was like, all right, well, what we're focusing on is people who aren't showing up.

Okay. So you do that one and you're like, uh, okay, well what does the inverse of that look like? Okay, great. Make that green perfect.

Evan Troxel: you know, speaking of the KM 2.0 days, which was the social side of things and the, and the likes and the shares. I mean, is that still something that's going on? Is that something that became cultural in and has stuck and, and, and now we're building on top of that? Because it seems like, I don't know, social media is, is a dirty word nowadays, right?

like, but, but there's a social aspect to sharing and there's a social aspect to the way people are consuming and talking about, you know, there's a buzz around certain things. I'm sure. I'm, I'm just curious what that looks like inside of Shepley on the, on a place like this where it's like you're disseminating information, but it really is a participation and engagement kind of a thing as well.

Jim Martin: I'd say it's still going strong. I actually did a post, uh, last on Friday of last week asking, uh, about informal mentoring. So just asking people, it's like, what questions should, uh, someone who's looking for mentoring ask someone who's further along in their career.

Um, and it was interesting, it got a couple of responses along the lines of, oh, you should do a 40 minute workshop.

Uh, do a conversation about that and. My response to that was to, uh, that came from people who are more senior in their careers. 'cause they couldn't think of anything to ask, which I thought was pretty interesting in and of itself. So I started poking every single designer and level one that we have in the entire company and asking them to contribute to that post.

And, you know, to date we've got 20 responses back on it of really interesting questions, which we're gonna turn into a Finch page and include example templates for, um, if you are a mentor and you want to offer to mentor somebody, you could use this, email template. And if you were a mentee who's interested in being mentored by someone else, you could use this email template. And then here's a list of 50 questions that you could ask, um, to give people inspiration for what to do. Just trying to reduce the friction, um, of the platform. And, you know, that is all social enterprise. I mean, that is. Anyone feeling comfortable kind of saying something. The added benefit was, is that people really appreciated being reached out to.

And several people commented that they thought it was a really important topic, um, to be thinking about. And it gave them an opportunity to be seen. And I'm gonna see if I can convince the CEO to reach out to every person who responded and say thank you personally. 'cause that's a nice touch at the end of the, of the effort.

So building campaigns like that, you can really use the social aspect of the internet in order to build a lot of positive change and a lot of, uh, tools and crowdsource things and get people thinking about, oh, I actually, hmm, I could reach out and ask someone, uh, for mentoring. And now I've got some thoughts of my own about it.

Christopher Parsons: But lemme just pause on, I mean. I hope people just like, so yes, there is a share box. You can write a post, you can get comments in the technology. That is 5% of the story Jim just told. Right? That

is a process culture orchestration. Why in the hell is Jim Martin, who works in IT, helping us figure out informal mentor mentoring on a Friday.

I have questions there, like outside the swim lines, but like, this is the kind of thing that I've seen Shepley do and not just Jim. I've seen Jess and other folks do stuff like this where it's a very, I feel like it feels secondhand to like, I don't, like you knew. That's like I need to back channel with some people

to get, it's like some, some voices heard, oh, I'm not just gonna leave this in the comments of a post.

We'll turn this into an asset. We'll figure out how to put it into the flow of work. Like that was such an elegant example of what like great knowledge management looks like. Whether or not you have knowledge management in your title.

Like that kind of thoughtfulness. And I'm curious, like I I said culture of experimentation, that's just like, that feels like a very lightweight. Like, maybe nothing would've happened. Right? Oh, well, you didn't put a lot of

like, you're

Evan Troxel: it out there, and then just

Christopher Parsons: Yeah.

Yeah, But like you, as you started seeing, you're getting some value, you started ramping up your effort, right? To see if like, oh, this has got some legs. Like, let's see if we can make something out of this. But really, like, Jim, why informal mentoring on a Friday?

Are you asking that question?

Jim Martin: Well, I, I was in the studio on Thursday and was talking to one of the directors

and we had a conversation and informal mentoring came up and it became clear to me that she wasn't quite sure how to go about it. And the process that I just articulated is a process that we, uh, had developed maybe six years ago or so, and we've got it documented.

We know what exploring out loud looks like and how to use the intranet and the culture and the willingness to be vulnerable that we've cultivated in order to generate value for the business. And, you know, that's how, so you basically, you come up with an important question that needs a whole bunch of people to kind of ruminate on and to think about, and you create a process that encourages that sort of stuff to pop up on there. Has a lot of ancillary benefits of connecting the firm together. We are hybrid, so getting people to connect with each other is really important because a lot of times firms use their physical environment in order to facilitate that connectivity. And at Shepley we don't have that as much. We have offices for certain, but not everyone goes into them all the time. And we want people to be as connected as possible because when they start working on projects together, they've got preexisting relationships to build on in order to make the project process easier. So there's a lot of things that kind of go into it, but it's, it's, it's understanding your tools and then understanding the way that you can use your tools in an elegant way in order to get multiple value propositions satisfied. So like, you know, I legitimately think that if, if people at our firm felt like there was more friction in order to reach out and get mentoring the firm would be a better place. But I also think that getting a whole bunch of, uh, early. Uh, career architects and interior designers engaged in thinking about this sort of thing together on Finch is also good for the firm. It's also good if we can create a content after the fact that is more curated, that has expectations around social norms kind of baked into it, um, that help people understand what is socially acceptable for them to do. And that, yeah, it's actually okay for you to reach out to your project manager if you do it in this way. And there's a, there's a bailout button baked into that process for the project manager. So we're not committing our future project managers is something they're not comfortable with, and we will kind of go through all that. So it's, it's all about developing process, but developing it together and not just being one person's.

I mean, it would be totally inappropriate for me as the CIO of the firm to sit in a room and think, well, here's a process that we will do that will help this thing. Bam. Done. You're welcome people. It's your process, but doing it with everyone else. I think it is. That is how I justify going out of my lane so

Christopher Parsons: It even

seemed like if you'd just gone to the most senior people, that wouldn't have worked either. Right.

Jim Martin: they're not the target just going to the, to the users. What's that?

They're not the target They're not the target audience.

You know, the target audience is the early career architects, interior designers that we have

because they're the ones that need the mentoring, whether they feel comfortable asking for it or not.

Christopher Parsons: Jess, I'm curious, as you're looking forward to 2026 and kind of the future, like what are you, what are you excited about? Like the, around knowledge management or learning or AI or any of the kind of things in that bucket? Like what are you, what are you dreaming about?

Jess Purcell: Oof. You always get me with a question that I don't have a good answer for.

Jim Martin: I know what I'm excited for, Jess about.

Jess Purcell: Yeah. You tell me and then I'll repeat it.

Jim Martin: We're excited about getting the ability to see, uh, all of our data with ai.

Jess Purcell: Oh yeah, I'm doing that.

Yeah. So I am currently moving, or at least copying, making a copy of all of the data we have within our environment inside fabric into one Lakehouse. And this is gonna enable teams, like operations teams to do easier reporting on it, but that also puts it in a place where it's accessible for AI tools.

So we're playing around with agents that can do things like if we wanna stay in the camera realm, I wanna build a database of metadata around our content that then we can use to make smarter decisions. So rather than the kind of traditional route with an LMS of like. anybody who's about to be a PM gets these courses or anybody who's a level one just starting gets these courses. You know, could it be something like, we have these, um, baseball cards, which is where people can self elect, uh, their skill level on different, uh, tools and things that you would do within design. Can you then have an agent that looks against an individual's skillset and then against all the content? And so you could even do something like, oh, this person has some extra time. What could I assign them? We have a cohort that's starting to do some specific thing. What courses do I have that I could point them at? You know, much more tailored individual assignments and identifying opportunities. Um, even like, I would love to get to a point where it could recognize quality of content. You know, how many pauses there are. Is like the audio bad? Uh, are there a lot of side stories and questions? There's a conversation in the LMS beta group around recorded sessions, watching them later. You feel like a second class citizen. 'cause there's probably only like 30 minutes of good content in there for you, surrounded by stories and questions and answers.

And those are all really great things live that makes a great engaging session live, but it makes a bad recording.

Christopher Parsons: Mm-hmm.

Jess Purcell: So can we use AI to start identifying chunks to pull out of a video to build into a course or remake into a course? Um, those are the kinds of things I'm looking at with AI within KM.

Christopher Parsons: That's awesome. And I mean, this is why we love working with Shepley too. 'cause I could take 70 or 80% of that, and I'm like, well, how can we generalize that and build that into product? Right. So

it's fun to watch people. I mean, Jim, you laugh, but like I think that's, you're cool with that.

Jim Martin: Yeah,

I'm

Jess Purcell: that's our

Christopher Parsons: You're like, how do I make something compelling enough that Chris will take it over and like we don't have to continue to support

it

Jim Martin: Amen.

Evan Troxel: Take the weight.

Jim Martin: I got

Jess Purcell: absolutely the strategy. We experiment to make, uh, basically a working proof of concept

that fits within the culture and solves the problems that we need.

And then sometimes we have people who just take it over. Sometimes we pay people to make it better,

but that frees up our time to try the next thing because we're not continuing to maintain internal tools.

Jim Martin: Yeah,

and we're not, we're not Gensler. We're not AECOM. We don't have a lot of resources for internal development, so it's really one of the only ways that we can get that the tools that we need built, built is to partner with people. And so we try and be really good at being a partner to people that we think can develop tools and software and processes that will help us.

Christopher Parsons: You guys are a good partner at that. I will say that I, the, the, the, um, I don't remember where I heard this quote, but it was, you can't innovate without innovative clients. And

I couldn't agree. I think it's just like, and it was, it was not in a technology context. It was an architect actually saying, and I can't remember who said it, but I was just

like, that's the one. Um, Jim, as you look forward, what's your hopes and dreams?

Jim Martin: Well,

riverside_evan_troxel_raw-synced-video-cfr_trxl_0297: maybe Jess should answer this one

Christopher Parsons: Yeah. Jess, what should Jim talk

Jim Martin: oh, yeah. All right. Excellent.

Jess Purcell: Yeah, that's a good question.

Jim Martin: Yeah, I would say for 2026, um, I'm also, I, I'm looking forward to the AI tools. They're infuriating to figure out, but there's so much value, uh, to be had if we can figure out the appropriate application of them and the guardrails that need to be in place in order to make sure that they're reliable, repeatable, and don't lie to us constantly. But the more information we have available, uh, to kind of do that, I think the more tools we can start to build, uh, that will help people solve a lot of the daily grind problems. A lot of the friction, a lot of the sand in the wheels sort of thing that people deal with on a daily basis. And, uh, I just think that we will have even more things that we can think of, uh, to apply it to that will add even more value as it gets better and we get more facile with understanding what it can and cannot do and what all the new stuff is.

Evan Troxel: I am curious what you guys would both say. Maybe Jess, we can start with you, just what would you say to firms who don't have any kind of knowledge management system in place? Or maybe it's. Totally ad hoc, right? It's, it's email, it's instant messaging, it's file servers, it's PDFs, it's all, you know, models.

It's all of those things, these disparate things that every firm has had deals with. You guys have been doing KM for a long time. If you were to go back and, and tell yourselves, like, this is where you'll be in 11 years, I bet you wouldn't believe it, right? So, so what would you tell people who don't have anything in place today who, who might be considering getting started with this?

Jess Purcell: I'm not convinced that no one's doing KM. You might, like Jim said,

not know that you're doing it. I mean,

architects are storytellers, so you have people in your firm who are sharing, you have people in your firm who are getting all the questions, and they want a place to answer that question once so they can get new, more interesting questions. So I think there's a lot of strategies we've built that we've talked about already around identifying those people that want to share. And making it easier for them to formally share. one of the things we do, uh, when we have live sessions of things and we want someone to do like a post on finish after, with like a recap and things they took away, I'll find the person who was the most active in chat, because then I'm not asking someone who like was maybe working on the side to like, think through like, what should I say about this?

This is someone who's already really invested in the topic

and asking them to just kinda like put their thoughts.

Christopher Parsons: I think Jess, like, um, from like a, from an outsider's perspective, I think there's two things that Shepley does that maybe you guys wanna expand on or not. But one of them is, you guys are the opposite of build it and they will come firms like you kind of already understand the culture, the, the change management, the process that has to go on behind the scenes to make a new thing sticky and kind of get momentum of its own. And I think there's this, um, you can't just buy a tool, you can't just write a check thing that I think is very clear from watching you. And I also think you're very clear that you can't manage all of the knowledge at Shepley. Like you said early on, Jess, like, we're not gonna repeat a Bluebeam guide, you know, and teach people how to use Bluebeam here, but we're gonna teach 'em how to, how Shepley uses Bluebeam. I feel like those are two things that you do. Really, your, your company does really well. That for somebody that hasn't like, invested either in knowledge management technology or formalized knowledge management, that would probably save them a lot of heartache is to understand that there's gonna be a process change management component, um, that goes on beyond launching a, I mean, you guys have been on Finch for 11 years and you're still investing in getting better at knowledge management. If, if that makes any sense.

Jim Martin: I think, uh, Carol Wedge, our former CEO was fond of saying that Shepley Bulfinch is 150 years old and a fixer upper.

Christopher Parsons: I remember that. Yeah.

Jim Martin: I think that is absolutely true, and the reality is, is that, uh, synthesis makes it easier for us to share knowledge, but it is not, it is not critically necessary for that process to occur. A lot of the things that we do are just, are agnostic of the platform. I mean, a lot of the behaviors that we built into people around sharing knowledge, uh, on synthesis evolved to do the same thing on, on uh, teams. So we have very active communities that have taken to using teams instead of synthesis to share really critical knowledge around the specific thing that they're talking about. Our sustainability group is on teams. Our, uh, Revit users group is on teams for two examples. Um, the problem is, is that if you aren't also on that team, you don't get to see all that knowledge. Uh, so there's, we struggle with that a little bit, but, um, I have made the determination that if we tried to force them to move, they would just stop.

And so I don't want to kill the golden goose. We just let them roll and we'll figure out a way to incorporate that back into the overall strategy at some point. But when people are doing something that you want them to do, uh, it's. I would say pretty important not to, uh, chastise 'em for doing it in the wrong place if you can help it.

Um, but different processes have existed with the firm over the years, uh, whether since this was evolved or not. Uh, all of the in-place workshops that we do, our knowledge management knowledge transfer around markets or around projects, we have project sharing and design and inspiration, related things at town hall every two weeks we have all these things that we do.

We do Shepley spotlights where Bulfinches pop up on town hall and tell a quick Shepley origin story or some other kind of related story, which helps people connect more effectively, which is one of the things that is important for us to do as a business. So. If you're worried that, oh, I don't have all the tools that I need in order to get started, get started with the tools that you have. Talk to some smart people who have already figured some stuff out. Uh, listen to some content on podcasts or, uh, other go to professional conferences that exist around the whole knowledge management, uh, domain. Uh, I still feel like I know very little about knowledge management. I know a lot more about IT than I do about knowledge management. Um, but I know some things that I can repeatedly, uh, use in order to make things that have value to our business. And they happen to also be knowledge management strategies. So anyone can do it. You're already doing it. Look around point at stuff, say that is actually knowledge management. And then make a list.

And you're gonna be surprised at how long that list is, I think, in your firm, even if it's not being done by a single person or, uh, a single department.

Christopher Parsons: That's great. I, Jim, you made me think of, I, uh, somebody on a podcast I was listening to recently, not, not in AEC, it was a, like a knowledge manager, AI kind of person. Um, was saying like, you could, I think you like, worked at McKinsey or something. He's like, we could take every knowledge management process that we have and we could take it back to paper. Right? You should be able to kind of like write, to figure out the paper version of whatever this is. We don't have to use technology to your point,

but then understanding that process at a paper level helps you understand how it works better, and now you can understand how to apply technology to kind of speed it up and scale it up or whatever your goal is.

Jim Martin: Yeah. I mean, in my example of the informal mentoring, we could absolutely just have an inter departmental envelope going around with a piece of paper on it, just asking everyone to keep adding to it. And then when we got to the end of the process documented and post, it

Christopher Parsons: Right. Yep.

Jim Martin: would've been a lot harder.

Christopher Parsons: I just think, I think that you're kind of, you know, process is something we've talked a lot about culture, experimentation,

probably the title of this episode and somewhere in the nexus of those words. But I just think I appreciate you, you two coming on to kind of underscore all of the non-technology aspects that

make, that make knowledge management sing. Um, because it's most of it. Um, and then, you know, you try and pick good tools. We try and build, make good, uh, build good tools, but, um,

Jim Martin: tell you, I've, I've said this I think to you before, Chris. The, the, for me, the software that knowledge management creates is fantastic, but not the primary value of Knowledge Architecture. For me, the primary value of Knowledge Architecture is the community of people that you have kind of created around the product that are all trying to do the same thing at the same time.

And it is so much more enjoyable to be on this journey with a bunch of smart people from a bunch of different places than just trying to solo it. So, kudos to the community and kudos to Knowledge Architecture for creating said community.

Evan Troxel: I think a big takeaway for me builds on what Chris was just saying, and it really has to do with the people aspect of all of this. So you've got the people in the process and the tools, and I, you know, you, you may have been kind of intrigued early on in your career. I know I was with, with tools and what was possible, and then cons that just, that just has built and built and built.

But to constantly go back to why do we do what we do and how do we facilitate that to happen at the most basic level and all the other levels of alignment to values that happen in between is super impressive. And I just want, Jess in particular, Jim, how lucky are you, right? To have somebody like Jess who thinks the way that she does about, about these things.

I think that's, it's absolutely incredible.

Jim Martin: Yeah, I am very

Evan Troxel: lucky. and then just your leadership and I, that's the other word that keeps coming to mind is just leadership with this. Um, in order to ensure like, like the intention behind all this is. So amazing. Right? And, and to, to keep your reputation of a firm of 150 years takes diligence to, to just stay on top of this stuff and roll it out with intention and make sure people are getting what they need because they are the engine behind.

It's not the tools, right? It's, it's those people. So I'm, I'm super, super impressed with everything that I've heard today. So kudos to both of you and the whole team at, at Shepley Bulfinch for really sticking to it. I mean, that to me is, is really the glue, is, is this leadership team and, and this constant looking out for the benefits and the needs of others.

I think that that's just super, it's really impressed upon me.

Jim Martin: Thank you. more, one more into that. This is for both Jess and Jim equally. I maybe a little bit more just 'cause of recency bias. 'cause we spent time working on our KA Connect talk together. But I think another thing that people can learn from the Shepley example is around, um, kind of what I was saying with the Jim piece is like going to end users.

Christopher Parsons: Like they could be interns, they could be job captains. Uh, a lot of times, you know, enterprise software like a synthesis, it doesn't get bought by end users. It gets bought by leadership and management and then kind of like people are told this is what you're using now and you only get so far with that, with that approach.

And I think what's great about what Shepley does is they go to the people, whatever the pro, whether it's a process, whether it's a tool, some kind of change, and they go to the people who are gonna be impacted the most and learn from and with them. To help shape the way that they message something, the way that they support or roll something out, and I think folding in those actual end users and make 'em the center of the story is something I think a lot of firms could learn from.

Jim Martin: Thank you.

riverside_evan_troxel_raw-synced-video-cfr_trxl_0297: Well, thank you all for taking the time to have this conversation today. There was so much wisdom and experience shared in this episode, so appreciate you. We'll have links to everything we talked about in the show notes for this episode, so be sure to check those out and that will include links to connect with both of these.

On LinkedIn and follow along and see what they've been up to. So again, thanks everybody. This has been a fantastic conversation.

Christopher Parsons: Thanks,

Jim Martin: Thanks for having us.

Jess Purcell: Thanks.

​