DEEP DIVE | KM 3.0: Connecting People to Knowledge and Expertise in the Flow of Work

Hey everybody. Welcome to the webinar. I'm Christopher Parsons, founder and CEO of Knowledge Architecture and I've spent the last 25 years helping AEC firms manage knowledge more strategically and effectively. And I'm excited to share some of our thoughts on where knowledge management's headed with you. I started my career right here in San Francisco running IT for two architecture firms in the city. Worked for those firms for about eight years. Combined with, and in both roles, I built intranets and launch knowledge management programs. And I did that from scratch, right. And my feeling was it was harder than it should have been. And so that's a lot of why we got Knowledge Architecture started is to help AEC firms do intranets, do knowledge management do knowledge sharing much more effectively. So when I started, there were no good intranet platforms designed for AEC firms. There were no books, there were no conferences or communities focused on knowledge management or industry. I had to do it on my own. And that's partly why we shaped and founded Knowledge Architecture, right? We built synthesis to provide the Internet and search technology. I wish that people had we started K connect in our community to build the community that I wish I'd had. And that journey, and really over 25 years at this point has led me to believe that we're at a transformational moment in knowledge management in AEC. KN 3.0. It's what I've been calling it, it's in the title and that's what this talk is all about. So I'm going to explain what I mean by km 3.0 in just a minute. But I want to introduce Knowledge Architecture for a little bit for people who may not know about us. So synthesis is our product and it broadly does three things. Number one, we help AEC firms build a searchable firm wide knowledge base. This could be standards, policies, procedures, best practices. We surface knowledge about employees so who knows what, who to go to for what, as well as key information about current and completed projects. We call that essential content. Number two, we support internal communications, so sharing between offices, between departments, leadership updates, project milestones, new hires, promotions, all the things. So we want people to stay connected to what's happening in the firm. And number three, we help firms become learning organizations or support them if they already are learning organizations. So this might be supporting onboarding or professional development programs and supporting questions and answers across the firm or lessons learned. And later this year we're going to be adding a learning management system into synthesis. What's new? So we've been historically been an intranet Platform, as I mentioned, we're adding learning management and we also added AI search this year. And so AI search layers over the top of all of this to connect people to knowledge and expertise in the flow of work. And so we're going to talk about a little bit more about that. Integrations have been a key part of synthesis from the start. So we connect with the platforms that AEC firms already use like Deltec, uninet, open asset, AEC 360 and more. These integrations unlock kind of a rich dynamic environment where you've got really rich information about employees and data, employee data rather like roles, experience, skills, who to go to for what as well as project history. So what have we done before? Where, by whom, how much did it cost? What was the square footage, those kind of questions. So that structured information becomes even more powerful when layered into AI search, as I'm going to show you in the webinar. So what is knowledge management? So there are lots of definitions on the web for knowledge management. We have actually changed our definition of knowledge management over time as we've learned more about it and it's changed. But here's how I would describe knowledge management today. It's a practice of using people, process technology and culture to help your firm make the best use of its knowledge. Right? All AEC firms are knowledge based firms. There is a wide range of kind of intentionality and how strategic firms are about using that knowledge to, to grow better, to do better design, to engage employees. There's a wide variety of ways that you can use knowledge. And so it's really about the process of being intentional. So I look at the core, right, about connecting people to the right knowledge around connecting people to the right people and doing both of those in the flow of work. And so what do I mean by flow of work? So let's imagine you're in the middle of a project and you need to find a standard or a template or an expert. You shouldn't need to stop what you're doing and launch a big research project. You shouldn't need to spend two hours or four hours or eight hours trying to find the thing you're looking for. Oftentimes people will do that or they'll just give up and then go out to YouTube or the Internet, find something and just kind of or reinvent something from scratch. And so KM in the flow of work means you take advantage of all that organizational knowledge you have and you deliver it to somebody at the point of need so that they can keep working on their project. And not get totally disrupted. So what I think is that we're entering a golden era of knowledge management at aec. This ability to keep people and deliver knowledge and expertise in the flow of work, I think is completely transformed because of AI. And when people ask me how work is going these days, I tell them I'm having the most fun I've ever had in my career because I've been working in AEC and knowledge management for 25 years. As I mentioned, there have been a lot of ups and downs, like knowledge management's been hard. We've made a lot of progress as a company, as a community in terms of managing knowledge. But I think that this AI unlocks capabilities that we in knowledge management and AEC have been dreaming about for decades. So I want to take you through some of that in this presentation. Okay, so I've got four parts. Part one, I'm going to talk about the three areas of KM technology and AEC. The majority of the presentation is part two, which is looking at how knowledge management plus AI search unlocks new business value. Part three, I'm going to talk about what KM 3.0 wants from an AEC firm. So what kind of cultural or process change do you need to do to take advantage of this new opportunity that technology is presenting? In part four, I want to talk about what's next and where we think all this is headed. All right, so let's talk about the three eras of KM Technology and aec. And I'll start by just saying knowledge management is older than 30 years. But this is typically really, when you think about where this term started and how long people have been talking about it, we kind of look at the mid-90s. That's when you start seeing it show up in things like Harvard Business Review. That's where you see organizations like APQC get started. And talking about knowledge management, that's when it really entered the ethos. Obviously, as a species, we've been managing knowledge for a lot longer than 30 years, starting with oral storytelling traditions to cave paintings, to tablets, not the new ones, but the old ones, the stone ones. You talk about educational institutions, we talk about companies, you talk about libraries moving into technology and the kind of things we have now. So we've always evolved and grown and changed the way that we do knowledge management based on technology. And, and I think over the last 30 years, we've seen three kind of key shifts. So number one was kind of circa 1995 when the Internet exploded and the web exploded. So it transformed the way that we accessed information outside the firm, but also inside the firm. And I think what happened is AEC firms started pushing towards digitization. So they moved from things like binders and file cabinets and static documents into intranets and file servers and searchable databases. This era was all about getting knowledge online for the first time, right? It was making it easier to store, organize and retrieve knowledge. So this focus, KM 1.0 was on knowledge capture and it laid the digital foundation for everything that came afterwards. KM 2.0, in my view, was around knowledge sharing. And this started around 2007 when social media exploded. So Twitter, Facebook, LinkedIn and I was still in practice at my second architecture firm at this time as a chief information officer. And I was looking at what was happening with Twitter and Facebook and LinkedIn and thought, what happens if we brought those tools inside the company for knowledge sharing? Like, what could we do? We were a multi office firm, so how could we use this to connect the office? How could I get answers to questions across the company? We started building very rudimentary internally things like activity streams and hashtags and comments. And we were essentially building, when we launched Knowledge Architecture In 2009, we were building a social intranet and that's how we presented and marketed synthesis. When we started the company in 2009 as a social intranet, this was the. Of course we were going to do the KM 1.0 stuff, right? The knowledge capture piece, and have information about standards and policies and procedures. But this thing that was new was making it so that anybody in the firm could ask a question and get an answer or share what they knew, which was awesome. But with every technology change, there's also cultural change that needs to happen. And there was a lot of cultural resistance to this idea of social knowledge, right? And sharing using social media tools inside a company. You know, people would ask questions like, well, what happens if somebody posts the wrong thing or says the wrong thing? What happens if they say something inaccurate? What happens if they say something inappropriate? This means that anybody can post without approval. And so many firms were hesitant to. They wanted us to put controls on like comment moderation and posting controls. But if you flash forward to 2025, you know, almost 20 years later, you know, with teams and Slack and products like Synthesis, it's hard to imagine running a firm without knowledge sharing tools like this. But at the time it was a big leap. And I think that's where we are now, right as we enter km 3.0, what I would say is the AI powered knowledge management era for us. That means AI search. That means video transcription. We're building a search agent. And if you don't know what agents are, don't worry. That's probably a lot of people now. I'm going to talk to it kind of in the next what's next section at the end. But my point is we're just at the beginning of this transition into the KM 3.0 era. It's the first inning, right? ChatGPT exploded in 2023. And we've all been looking at how can we take advantage of this technology and use it inside the company. So similar to the KM 2.0 era that I talked about, there's a cultural shift for KM 3.0 that's going to be required, a process shift, and not just in tools, but how we work. And so I'm going to talk about that in part three. Okay. So I just wanted to set the foundation for what I mean by KM 3.0. I wanted to talk about the fact that AI search and what we're tackling with AI powered KM builds on what we did before around knowledge capture and knowledge sharing. And by we, I mean this community in this industry, not just knowledge architecture. All right, so part two is built around five case studies of AEC firms and how they've entered the KM 3.0 era and what kind of new business value that they've hoped to unlock and have been unlocking with these new tools. So we're going to talk about BWBR first. They're an architecture firm based in St. Paul, Minnesota. They're about 200 people, multiple offices. A key challenge that they're facing is that they hired 70 new people in the last year. Many of them are working on healthcare projects, which I think many of you know are notoriously complex. There's all this regulation and there's nuanced building types. There's tons of acronyms and jargon and specialized processes and equipment and room layout stuff and user types. It's a steep learning curve even for an experienced AEC professional, let alone somebody who's new to the industry and new to working inside an AAC firm at all. So the KM challenge at BWBR was how do we get 70 new people up to speed quickly in a deeply specialized knowledge area? And so I'll show you how knowledge management 1, 2 and 3 are making a big impact there. So BWBR's intranet is called Beehive. Most of our clients rename them. So you'll see in the upper left, it's Beehive. And so what we're looking at is a healthcare community of practice page on Beehive focused on healthcare. So community of practice, if you haven't heard the term, this is a knowledge management term, is a group of people who share a domain. So it could be a market, it could be a building type, it could be even a technical skill or a piece of software like Grasshopper. It's people that come together to share knowledge and learn from each other. So you have experts who are kind of sharing what they know and answering questions. You have novices that are trying to learn all this new material. They can ask the questions and they can also provoke a lot of change. So keep. Communities of practice are an old thing. I would wager that all of you on this call have been in a community of practice at one point or another in your life, whether you called it that or not. So this is a healthcare community of practice page. One of the things that that community of practice has built is a healthcare upskilling training series. So you can see Here there are 18 modules on their intranet covering foundational topics like Medical Planning 101. There are courses on emergency departments, clinics, surgical services, inpatient units and so on. So this is really trying to cover that core knowledge needed to work effectively in this complex building type. This is what BWBR calls a practice hub. So it contains a ton of best practices on doing healthcare work, including this example, which is a page on medical imaging, which is a highly specialized healthcare area. And so you can see that there's curated content and in the bottom right, an expert to contact if you've got a question. So this is like classic KM 1.0, capturing and organizing knowledge so other people can find it. But here's where things get exciting. So we've layered AI search on top of all this content. So the AI doesn't just do keyword searching, it understands questions and it pulls the most relevant answers from across the Internet. So long form pages, videos, documents, social posts, et cetera. So instead of having to know where to look or who to ask, you just ask a question. And AI services the best pieces of knowledge across formats and time. So let's say I'm an emerging professional, I'm working on a healthcare project and someone mentions or I get assigned to work on an onstage offstage layout and I have no idea what that means. So I go to synthesis, I type in the question and it searches through all the content on the Internet and generates an AI generated search summary. It's telling you exactly where it pulled that information from you can see the citations kind of at the end of each of these paragraphs. So in this example it's citing a video from that upscaling series that we saw before in that community of practice on ED planning concepts. It even tells you the timestamp. So it's minute 126. So if I were to click on that video, it, it would take me straight to that part of it. Or I'm an emerging professional, I'm hearing this term fgi. And if you're in healthcare, you know what FGI is. If you're not in healthcare, you probably don't. So what is this term? When do we use it? What is it? So I type in what is fgi and when do we use it? And there I've got a generated search summary from content that we've seen before. An example, here's a document coming from one of those learning modules the that we looked at. Or I'm an emerging professional, I've been asked to start working on an imaging suite, but I've never worked on one before. So I might ask what's important in an imaging suite? Synthesis generates an answer. And there are eight things here to consider. And one of the things that I think is interesting here is if you look a little bit carefully at the end of each of these pull ups, each of these strategies or factors, you'll see that it's synthesizing knowledge from different people at different times in different formats, all coming together to put the best answer that the firm has on that question. So for example, the first citation is from one of these best practices pages on medical imaging. We saw that in the Practice Hub earlier. Here's a question on. I'm sorry, this one is coming from a video. It's coming from. You can see 1 hour and 14 minutes into this video, on this video on hybrid operating rooms. I just want to pause for a second and talk about what this means, what KM 3.0 means. And I want to take you back to the old days of search. And by that I mean like six months ago you would have done a keyword search on imaging suite layouts and probably gotten back like 12 documents and 10 pages and seven videos kind of in a grid on your, on your page, right? Each of those videos is 30 minutes, 40 minutes, an hour, maybe longer than an hour, like this example. So what is an emerging professional supposed to do? Like read all those pages, read all those documents, watch all those videos, like seven hours of videos to find an answer? We know that's not gonna happen. Right? So we've done a good job historically in knowledge management of capture and sharing, right? We've got really rich content, we've shared really rich knowledge, but where it's fallen down, and this is why I think there's such a kind of golden era coming, is on retrieval. It's when somebody needs that piece of information in the flow of work, right? It was harder to retrieve than it should have been. But now this kind of AI search in this KM 3.0 era is changing all that. So it lets me as an emerging professional, stay in the flow of work, get my answer and go right back to my project and start working. I mentioned before in my definition of knowledge management about connecting people to knowledge, but it's also about connecting people to people. And that's a big part of what we're trying to do in this KM 3.0 era. So thanks to those integrations I mentioned before, we have a lot of structured data about people, titles, bios, skills licenses, certifications, but we also capture insights about expertise via that intranet content. So I was pointing out those contact blocks on those pages before where they said this is the go to person for imaging suites or talk to this person about NICU design. So our AI search looks across both the structured content and the unstructured content to answer a question like who should I ask about NICU design? And just like that, you'll see it identifies two people in the firm who are the most likely. Go to experts. Or maybe I'm asking who can help with infection control design and synthesis. Comes back with three experts. And right here I've got a link to Devin's profile so I can click on it. I can understand what she's worked on, get a sense of her background. From there I can reach out to her on teams, ask for help, whatever it is that I need. And so this is exactly what we're trying to do with AI. Not just connect people to knowledge, but connect people to people and find the right person, build a relationship, kind of level up faster. And so when we're thinking about upskilling, especially in specialized areas like healthcare, this is a really important workflow to enable. So I just want to do another one of those flashback six month things, except thinking about connecting to experts. So what did it take to find the right person before AI search? Right. I might send a firm wide email and hope somebody replies. Or maybe I post in teams or Slack and hope it reaches the right person. Or I walk around the office asking, hey, do you know who knows about infection control design. It's slow, it's disruptive, it's hit or miss, especially in multi office firms. You compare that to what we're looking at now is who can help me with infection control design. I just found Devin and from there I can connect with her and get going. So I'm seeing like a whole new this is why I said again I'm thinking this is a golden era because this is a whole new way of working. All right, so that was case study 1. We looked at upscaling emerging professionals. I want to talk about Turner Fleischer and I want to talk about driving technology adoption at scale. So Turner Fleischer is a 250 person architecture firm based in Toronto, Canada. They've been very design technology focused for well over a decade. So they've pushed tools, they built tools, they built a design technology department. They're very tech forward. But, and I think especially the technology people, but anyone that's done change management can relate to this. It's one thing to have a design technology team building cool tools, knowing about cool tools, introducing cool tools. It's a whole nother thing to have those tools adopted and used on progress and to drive that change and that innovation across the whole firm. That's the business goal that we're going to Explore. From a KM 1.0 perspective. Turner Fleischer has done a really nice job of capturing knowledge. Here's an example of from their digital practice page that really kind of gives a high level overview of the different resources that are available, whether it be about computational design or content creation or revit standards or visualization. If I drill into revit standards I can see in the upper right I've got key contacts to talk to about revit standards and then I've got a list of all the different resources that are available to me. If I drill in further into something specific like door family standards. You can see this page is structured around door parameters and different hardware schedules and door swing detail level. And it's just a very deep intranet with a lot of knowledge. And again key contacts. In the upper right, synthesis still includes a lot of that social intranet technology I talked about before that we added during the KM 2.0 era to support knowledge sharing. So an example from Turner Fleischer is they run a series called Tech Talks. These happen monthly, they get recorded and shared on the Internet. So this could be a new technology, they built a new technology, they're introducing to the company trends that are happening in the industry and they're recorded and shared on the Internet. So you're seeing all the posts, you know, with a hashtag TechTalks. So it's a great way to make this informal learning visible and accessible across the company. This is a video library on Newton. That's the name of their synthesis Internet called Etutor. Right. So you can see there are 374 of these little short videos, most of them around four to 10 minutes long. Each one focuses on a very specific task. So you'll see there are 10 videos here on ACC, the Autodesk Construction Cloud. There's 11 on Bluebeam and so on. So these are a great example of what's called microlearning, right? Bite sized, just in time content that helps raise tech literacy across the firm. And I should point out, in the KM 3.0 era, all these video transcripts are created automatically and are now searchable via AI. They also have a super robust learning and development program they call TF Academy. So what you're seeing here are the courses being offered in the current term. So in addition to that microlearning on demand library we looked at, they're also running more structured in person cohort based courses too. So they're putting a ton of effort into educating the company on what it needs to know about these Design technologies from KM 1.0 to 2.0. So documentation, social sharing, microlearning, in person courses, it's a wide range of approaches. They built a really solid foundation on these KM 1.0 and 2.0 strategies. So it's a great way to scale this expertise beyond just the DT team. So how has that set Turner Fleischer up for the KM 3.0 era? So imagine you're new to the company, you keep hearing this term BXP fly around, or maybe you've been at the company for a while. But this new thing gets emerged against introduced bxp. So you're wondering what's a bxp? And you feel you should know, but you don't. And you don't really want to ask anyone because it seems like they already all know. So you just look it up, right? What's a bxp? And you find out it's a BIM execution plan. And synthesis tells you not only the definition, but like how we use it at Turner Fleischer. And then in addition, when we have that information available, we'll also include a note at the bottom that says, you know, go to Brent Moudi for more information on BIM execution plans. So you not only get the definition, but kind of the clear person to contact to. Okay, so now I've been told, hey, you need to go capture existing conditions on this new project that we've got. Cool. How do I do that? So I type in the question and AI search gives me a clear direct answer. And if you look at the bottom, oops. If you look at the bottom, you'll see a link to the laser scan request form. So I can take action right away. Or maybe I have to create a door schedule now. So I get back a bunch of different resources that help me, guide me how to do it, including a video that comes from that Etutor library that we looked at before, the Micro Learning Library. So whether I want to kind of read or skim or watch, like, I've got a bunch of different resources here that are available to me. Maybe I am kind of using Grasshopper and I kind of gotten going and I've gotten a little stuck. And I'm looking for someone who can help me so I can ask who's the go to person for Grasshopper? Great. Or maybe I have to use the laser cutter now for the first time with a project. And I've kind of watched the video, but I don't necessarily understand, you know, how to do it. So boom, we give you the two people that can help you with the laser cutter. So I think this is an important point that, you know, I think Turner, Fleischer and a lot of firms have done a really, really great things, especially when they're doing innovative work. You've done great training, you've documented how things work, you've shared knowledge, but then it's two weeks later or two months later or two years later. Right. And they don't necessarily remember everything from the training because they're humans. Right. And that's where just in time knowledge becomes really important. So when you go to a class, for example, it's oftentimes just in case knowledge. It's kind of covering a lot of broad topics that help you understand context and know what you might need to know one day, which is good. I think we should do that. Right. It lets people understand the art of the possible. But when someone actually needs something to apply to a project today, we also have to do just in time knowledge to make it available to them. Right. In the flow of work. So that's the power of KM 3.0 in our view. Okay, so case study three, we're going to talk about Boulder Associates. They're a 250 person, architecture, interior design, and Consulting firm based in Boulder, Colorado. Their intranet is called Biscuitnet. For reasons I do not have time to explain today. At their core business challenge we're going to look at is this. How do you scale expert knowledge across the firm? On the left is Todd Henderson, Director of Practice Improvement at Boulder. And on the right is Darcy Hernandez, who is a principal at Boulder Associates with deep expertise in health care. Like many of your experts, she's super busy, highly billable, high bill rate, and she's in demand, both by her clients, but then also by colleagues at Boulder. So Darcy has this kind of critical knowledge around this California legislation called SB 1953, which is requiring seismic compliance for healthcare projects by 2030. It's a huge issue and many of the healthcare systems in California assume the deadline would get shifted, but it hasn't and it's sticking. And so it's creating this urgency in California around getting a lot of buildings up to code really fast, which is real challenges for Boulder's clients. Right? There's financial pressure and logistical complexity and all kinds of trade offs that healthcare systems have to make. So Darcy is deep in the weeds with understanding how this, not just what's published publicly about this bill and the compliance, but what's going on behind the scenes to really make this happen politically, organizationally, et cetera, within these healthcare systems. So instead of asking Darcy to write something and kind of just like share her knowledge, Todd on the left, who's the intranet champion and KM Champion, interviewed her on Zoom, right? He recorded the video, uploaded it to synthesis, which then automatically transcribed it. He then took that transcription out to Perplexity and ChatGPT, kind of external AI tools to write key takeaways and summaries from the video, posted that back on the intranet and now he's made that available to AI search. So, so it's kind of this full AI trifecta, AI to transcribe, AI to kind of summarize and build the page, and then AI to retrieve as part of search. So the AI part of this, I think is undeniably cool. But the part that's really interesting to me, even more interesting perhaps in the AI piece, is that I said before that knowledge management was about people, process, technology and culture. And this process that Todd has pulled together is really, really powerful because asking someone like Darcy to sit down and write down everything she knows, it's unrealistic. Number one, she doesn't have the time. And number two, it usually doesn't work like it's not effective. We're Trying to capture deep, tacit knowledge from Darcy. And that's why expert interviews like Todd did are such a powerful technique because experts often don't know what they know. They can't imagine what it was like to be a novice. They, they don't know what they know is special. And so somebody that can act like Todd and be naive to the topic or pretend to be naive to the topic can really extract really deeply buried kind of insights and assumptions and create really useful content. And it's also efficient for the subject matter expert. So Todd took this whole AI trifecta approach of the interview transcription kind of summarization using AI and then making it available for retrieval with Emily Nightingale, who is another expert at Boulder. She has deep expertise in Bluebeam, especially around its use in project management workflows. So like Darcy, she's busy. It's hard for her to write down documentation from scratch. So Todd interviewed her, did the transcription, the whole thing. Super powerful. Okay, so what happens now? So I'll search on SB 1953 and I get a good answer that's citing the knowledge created in the interview that Todd did with Darcy. We can see that we're citation and we can point right back to that interview. It's also saying, hey, if you need further information, you may need to contact Darcy. So if you read this page but you still have a follow up or a more complicated interpretation of this, this is a good way to do that. How can we help clients comply with 1953? So not just like what is it? But like, okay, so what are we doing to help clients? And there we go. We've extracted information on that topic straight out of that expert interview that Todd did with Darcy. Or same thing with Bluebeaming. Right. So Bluebeam redlining, this is from that interview that Todd put together. Or if I want to know when should I use Bluebeam sessions versus projects again, extracted out of that interview and it's comparing sessions and projects and telling you the pros and cons of each. Or I want to see who our Bluebeam experts are within the company. And there's our friend Emily Nightingale right there at the top. So again, whether I need the knowledge or to find the person behind the knowledge, this kind of KM 3.0 AI search approach helps me do both. All right, so I've got two more case studies. The fourth one is on Bora. Bora is a 70 person architecture firm. They do interior design and planning. They're based in Portland, Oregon. Their intranet is named Chaco and their kind of focus has been on using knowledge management strategies to help scale expert knowledge. So this is Corey Squire. He's Bora's Director of Sustainability. He is a nationally recognized expert. If you don't know Corey on sustainable design. Before Bora, he was the Sustainability Director at Lake Flato. And in between those two things, he was a sustainability consultant. So he knows a lot about the subject. He was actually the lead author of AIA's Framework for Design Excellence. And in 2023, he published a book called People Planet Design. If you're looking for a great book on sustainability, I highly recommend it. There's a bonus in here that there's a whole chapter on knowledge management. And the reason why there's a chapter on knowledge management in Corey's sustainability book is his view is that knowledge management is the key to unlocking sustainability. So he would say that it's not about shaming people into doing the right thing. It's not about who cares more about the environment or the penguins or the polar bears. It's about sharing practical knowledge that helps people make better design choices to make buildings perform better. And like many experts, Corey doesn't want to answer the same basic questions over and over again. And that's honestly part of why you write a book, right? To get the 101 and the 201 level questions out of your head so that you can move on to harder challenges and deeper work. So, on their intranet, Bora is using that same principle. So we're looking at the best practices community. On their intranet, we're looking at a project outcomes page. And it includes contributions from Corey and other sustainability and design experts within BURA on things like materials or toxic chemical avoidance or indoor air quality, embodied carbon, environmental justice, other goals that BURRA has and shares with their clients for their projects. So these aren't just technical checklists, right? They are the project outcomes that Bora wants to achieve through smarter, more responsible design. And, and so this is a good example of taking knowledge management as a way to operationalize values, not just say, we want to do, you know, have higher indoor air quality or we want to create more environmental justice with our buildings. But like, no, literally, here's how you do it, right? And so an example would be the indoor Air quality page. Each of these pages follows a pretty consistent format. It starts with why the topic matters kind of period, and that outlines Bora's approach, so how the firm actually handles it in practice. And. And then it breaks down the key decisions and design strategies you would use to create better indoor Air quality in your projects. This is critically. It's like how Bora does it. And you'll see, of course, that Corey's linked in the upper right and he's the go to person if you have more questions. Corey is prolific on his Internet when it comes to sharing. So he's regularly writing about new and emerging technologies, research and development experiments, useful trends, what other firms are doing. Kind of like what's happening in sustainability. So this is a good example of KM 2.0, right? That social collaborative layer. This is actually Corey's book. He uploaded it to the Internet, right? He broke it down into individual chapters and posted the whole thing to a document library. So Corey and I had been talking about AI search in KM 3.0 for a long time. When I told him we were going to start building this technology, we started asking the question, you know, what would happen if we could build a digital twin not of a building, but of Corey, Right. And we nicknamed it Robocory. And his son made this illustration. This is his dad, right? As a robot who can answer questions on the Internet. And the goal is not that Cory doesn't want to talk to people like that's not what he's trying to achieve with Robocory. It's just that there are tons of, like I said before, 101 and 201 level questions that Robocory can answer really well, which frees up carbon based Corey or real Corey to focus on the more strategic, complex and high value work the client, specific questions or challenges that are facing the firm and their clients. So it's not just about making Corey more efficient, it's about unlocking capability for the staff. So if Corey's traveling or he's at a client site or on a project or out on pto, the firm can keep moving, right? People can get their answers in real time. It's expertise on demand. So let's take a look at Roboquery in practice. So I might ask, what's our firm standard for indoor air quality? And here's the answer. And you can see that the performance requirements page from their process community is cited. So if I want to, I can go deeper and explore the source material or I'm looking at what are strategies for minimizing CO2 levels within the building. And here are six of them and they're all cited and they're referenced back to underlying content, including, including in this example, Corey's book, which is pretty awesome at the bottom of the page. Again, as we saw an example before, you can see that you could reach out to Corey. So maybe you've digested this information and you haven't found what you need, or it doesn't go deep enough in a certain area. Connect with Corey and he can take you further. Who should I talk to about Embodied Carbon? These are folks that you can reach out to. And maybe Corey's out of the office, as I said before. So we're also suggesting you connect to either Debbie or. Or Jeff. And then who are energy modeling experts. Great. Okay. So my final case study is Shepley Bulfinch. Shepley Bulfinch is a 200 person international architecture, planning and interior design firm with offices in Boston, Hartford, Houston and Phoenix. Their intranet is called Finch. And like every AEC firm, Shepley faces this universal challenge around onboarding. So the question is, how do you balance giving people enough information without overwhelming them? So during onboarding, there's a natural desire to share a wide range of information. Policies, benefits, org structure, key tools, firm history, culture. But new hires often struggle to retain it all. It can kind of feel like a fire hose of information, which you don't want to do. But on the other hand, you don't want to just throw them into the deep end and ask them to swim with no support and no context. So that's the tension we're all trying to navigate in all of our firms. Too much too early or too little too late. So Shepley has built a really strong library of onboarding resources. But the real KM opportunity, in my view, comes after someone is six weeks or three months or even a year in. So when they actually need that policy or that process of that person, how do they find it? So let's take a look at that. So this is a page on Finch called I'm New Here. And it's a page style that Shepley pioneered a few years ago. Several years ago, actually. At this point. Since then it's been shared within our KA community. And many firms on this call have adopted an I'm new here page. And the idea is that you give a new employee everything they need in one place. So orientation information links to different groups within the office, how to get started with technology, where to find firm standards, and so on. So it's super thoughtful and practical and it's a great resource for new folks. The next page is a benefits guide. So it covers the kind of things you'd expect to see during orientation. Medical and dental insurance, 401k, employee assistance programs, wellness, and so on. This page is called the ropes course guide, or this is the introduction to the ropes course guide. It's a series of seven guided discussions that introduce new employees at Shepley to a broad range of topics in architectural practice. Like, There's a Design 101 orientation. There's a page about team roles and design phases. There's stuff about construction information and client meetings. So it's a really thoughtful way to understand how Shepley does things. Right. And to tell people not just the what, but like the why and the how it's done at Shepley. There's a Shepley Standards guide. This is everything from file naming conventions to project management, design standards, technical documents, sustainability, and so on. Right. And again, it's not just the what, it's the philosophy behind. Behind the how, the why, behind the how. And in each example, they've got key contacts listed. There's a page about project roles, which break down all the major roles and what's expected of people. There's a staffing overview page which explains how people get assigned to projects. At the center of that, or at the bottom of that, rather, is the staffing responsible body. So this is the group of individuals who are responsible for making staffing decision. Working groups at Shepley are essentially voluntary committees within the firm. So you can see there are working groups on a design exchange, Interiors, revit, project management, sustainability, the technical advisory group. So each of these pages make clear what are each of these working groups, how can you get involved, who's in each group? And finally, this is the course catalog for Shepley's learning and development program called Birdfeeder. So there are recorded sessions covering topics like project management, design, technical skills, professional development, marketing, bd, et cetera. So it's a great way for new employees to get oriented on what's available, but then also start learning right away. Okay, so going back to the point I made earlier, Shepley's done, as you can see, an extraordinary job creating high quality onboarding resources. That said, we all know when someone gets four weeks or four months into their job, they start forgetting some of those onboarding details because they're human and they can only retain so much. So, for example, if someone wonders, do our benefits cover a gym membership with AI search, they can just ask the question and get an answer. And yes, we do, and it's cited from the medical and dental insurance page. Or maybe they started working at the firm and they were working full time in the office. Right. But now there's been some kind of life change and Maybe they want to kind of understand if they can work from home or maybe hybrid or part time. They don't know. And so they can just ask, what's our work from home policy? And then get a clear summary of how it works. At the firm. They didn't need to know this information when they're being onboarded, but now they want to know it. Or maybe expense policies were covered during onboarding, but they weren't on a project where they were going to submit expenses. So they didn't retain it. Why would they? But now it's six months later and for the first time, they need to figure it out. So they ask, how do I submit expenses? And they get detailed instructions. Or maybe somebody joined the firm and got put on a project right away and that project's been running for a while, but now it's wrapping up, they're starting to wonder, well, how am I going to get assigned to my next project? And again, they may not remember that process works and how it went in onboarding. So they ask, how do we get assigned to projects? And they get a detailed summary that pulls together the firm's process for assignments, including working with a staffing buddy, coming from frequently asked questions, or how to maintain your baseball card, which is a Shepley tool for skills and experience and those kind of things. That was shared in a staffing update in a recent tower hall at 22 minutes in. Or maybe you might remember we talked about that staffing Responsible Body. They just want to know who's on the staffing Responsible body. So they search for that and boom, you get an answer. This is filtering the employee directory and returning the 12 people who are part of that Staffing Responsible Body or who knows Affinity photo in Houston. So Shepley doesn't use the Adobe suite. They use a different product called Affinity. So let's say a new employee in Houston is used to Photoshop, but now they need to figure out Affinity. They can ask, who knows Affinity photo in Houston? And there we go. It returns Anna as someone with the right experience. And then I can just click on her profile, reach out on teams, walk over to her desk. Whatever it is, I can connect to the right person who can help me kind of move forward with this new technology that I don't understand. Okay, so that was the heart of the presentation. I want to wrap up with a couple thoughts on culture and process change and talk about what's next. So, as you might remember, at the very beginning, we talked about knowledge management, being people, process technology and culture. And I'VE touched on process and culture a little bit here, but I've mostly focused on technology. So I want to kind of talk a little bit more about people, process and culture and what has to change in order to take advantage of these new benefits that the KM 3.0 ERA is making available. So I've got kind of five things. One of them is curating high quality, firm wide knowledge. That's number one for me. And this has always been true in knowledge management. I want to say that by high quality I mean acquiring new knowledge that's important to capture, keeping it up to date and removing outdated or inaccurate information. And that's always mattered. But with what I've just shown you over the last 40 minutes or so with AI, you can understand how the ROI of this information has gone up, but also the importance of keeping it up to date and getting coverage on important topics is really, really mattering now. So it's a operational change to not just do one and done uploading of information. It's a process change to constantly be reviewing your knowledge to make sure that you've got the new stuff, that you're updating what's currently there and then you're archiving things that are out of date. The second thing I want to mention is what I'm calling moving beyond the what. When we're talking about emerging professionals like at BWBR or folks trying to understand design technology or healthcare or sustainability or how we staff projects at Shepley, it's important that people understand more than just what, but the why, the how, the who really providing context. So context is really important for people, right? They really, especially if you're new. Like when knowledge gets written at an expert level, they tend to write at an experts, they tend to write an expert level and they suffer from the curse of knowledge. They don't remember what it was like to be new. And so going beyond just what, like so for example, let's say that we wrote something that says our standard for window wall ratio on projects is 30 to 50%. Cool. When should it be 30%? When should it be 50%? That's a pretty wide range. Like how do I know this? Like so providing that context and explaining more about why the standard or why the best practice is what it is what some alternatives are some more context and obviously providing the context, providing context of who is important. So that context is all really, really useful for people. But it turns out it's also really useful for AI. AI. So it allows it to answer some of these more complex and difficult questions. One of the features I didn't show you of synthesis AI search, that's kind of a killer feature for us, is search, rating and feedback. So whenever somebody does a search on synthesis and they get one of those AI generated summaries, they can rate it from one to five stars, which is good. But the more important thing is they can provide feedback that goes written feedback, right? That goes to both the knowledge managers and also to us at Knowledge Architecture. So in the past when people had a hard time finding stuff with search, with synthesis and other platforms, the feedback to the knowledge management team would usually be like, search doesn't work, or I can't find this, can't find what I'm looking for. But now firms are getting really rich, specific feedback, right? This page is outdated, or I'm looking for something on twinmotion and I didn't find it. Can you please add it? Or this result was misleading because it only told part of the story. So you're getting such rich feedback. The important part of that is that, that knowledge managers take action, right? And so this virtuous feedback cycle can happen where feedback comes in, content gets improved, search gets better. More search, more feedback, more improvement. It's a beautiful flywheel, but it only works if you actually take action on that feedback and improve the content. I mean, that's the only reason the content gets better, is because you do something. But it also keeps the feedback coming. And in my experience in work, if you ask for feedback, whether it's a survey or it's on a platform or whatever it is, and people take the time to give you feedback and you don't take action on it, the feedback will stop coming in because they're like, what's the point? Nothing ever changes. So the cultural change, the process change for knowledge managers is to take action on that feedback and drive continuous improvement on the content. The fourth point is about prioritizing strategically. And the reason I say this is that there's more knowledge in your organization than you can ever capture. This has always been true in knowledge management, but it's especially true now. Like, knowledge is moving so fast, the half life of knowledge is shorter than it's ever been. There are too many tools, too many materials, and too many topics. Like, you can't capture it all. So you have to prioritize your efforts. This is why I walked you through those five examples, right? Each one of those was tied to a clear business goal at the company. That's where you want to start. You want to focus knowledge capture on the areas that will deliver real business value. So if your biggest opportunity is in upskilling healthcare teams, start there, get that into a good place, and then move on to the next most important thing. Don't try and boil the ocean and capture everything at once. Just move systematically through strategic priorities. And last, I call this play like an orchestra. So you've all probably worked at places where there was like one person tasked with maintaining all the intranet content or all the knowledge. This wasn't a good idea in 2005 and it wasn't a good idea in 2015. It's definitely not a good idea now. There's just too much knowledge in the organization. So it really involves orchestrating and bringing people into the flow. It doesn't mean every expert has to write. We saw the example of Todd interviewing expertise and shaping and sharing the knowledge. But what you have to do is engage really as many people as you can in maintaining and evolving and growing your knowledge base. Okay, so I'm going to leave off on kind of like what's next for KM 3.0? What's next for AI search and synthesis? And where we're thinking, you know, the next six months, the next year, you know, kind of beyond now. So we have 120 firms in the AI search beta, over 120 firms, actually. And of those a quarter or over 30 are already live to all their employees. So this has moved way faster than we thought it was going to, which is really exciting. And the technology keeps moving too. So I want to talk just for a second about how the technology works, because I think it's a really important thing to know. So when we went to public beta in the end of last year, we were using this process called retrieval, augmented generation, which we're still using. But what this means, it's called RAG for short, is that we're not training a model on the firm's knowledge and then asking it questions. We do a search first to get the best possible information, and then we pass that knowledge to a model to then summarize and generate the answer. So that's rag, right? We do retrieval and we use that to augment generation. The next evolution of that, and this is where things get really exciting, is something called agentic rag, or building search agents. So here's how it works. A query comes in, say, who can help me with a drone survey? What we do is we send it to AI, an LLM, a large language model, and we say, what is this question about what kind of search is this? Is this a General knowledge search. And, and if the answer is yes, it follows that top branch on the top of your screen. If the answer is it's more like an employee search, well then we've categorized it and we follow a separate branch. We call this our query tree. So then if it's an employee search, we say, great, you've got these tools at your disposal. You've got tools for searching by employee name or role or title or expertise, or filtering your directory. Which tool do you think you need? Okay, if it's who can help me with a drone survey? That's an expertise query. So it then will analyze the content on the Internet, it'll search the employee directory, both of them. If it finds what it needs, it'll return the answer. If it doesn't, it'll kind of go back a level, try a different tool, kind of loop through the different tools it has available to it to generate knowledge. So this is why this is called Agentic Rag, or this is why people are calling it search agents. It's an agent that's acting on your behalf. You give it tools, we give it knowledge and we give it a goal. And, and we basically say, go navigate and find us the right answer. So this is what's next for synthesis. Agentic Rag for employee search is going out to all of those 120+ firms in the AI search beta next week. And then Projects is the next major capability for AI search in Nigentic Rag. And so we're adding a branch to the query tree for projects that looks very similar to what I just showed you for employees. And here are some screenshots from the lab of what we're calling Advanced Project Search. So I searched on what higher ed projects have we built in California, searched the synthesis project directory, and this data is sourced from systems like Deltec, Uninet, AEC360 and Open Asset. Or maybe just to show off a little bit more, what higher ed projects have we built in Massachusetts which were over 200,000 square feet? You can see kind of in the search summary at the last sentence, these following two projects, you know, match the filter project tape type equals higher education, contains the state Massachusetts, and square footage is over 200,000 square feet. Neat. Please list our LEED silver hire projects and you'll see in the search summary that we figured out to include silver, gold and platinum projects. Or finally, what projects have won AIA awards. So this is pretty cool. It's in development now, it's going to be heading into beta once it's ready and No, I don't have a release date yet, but we'll keep you guys posted as we move forward. And my last slide, and this is something completely different. Last part of the presentation is we're building a learning management system into synthesis. So we've been a intranet platform since the very beginning. For as long as I can remember, people have been trying to make synthesis into an LMS because we've had video tools and a great page editor and search. They've tried to cobble together learning management inside synthesis, but we haven't had assignments and transcripts and continuing education credits and those core workflows and analytics and structure you'd expect from a proper learning management system. So we've decided to build one and that's what we're working on now. It will go into public beta at the end of this year. And we wanted to do this for two reasons. One, to help our clients improve learning and development. Kind of full stop. Like we just want to move into learning development more. More intentionally. But also because when we greenlit the LMS project, we also knew we were going to build AI search. So if you think about what AI wants, right, high quality, firm, wide knowledge, you definitely find a lot of that in an lms. So now with our integrations, our intranet content and learning content all under one roof, with AI search layered on the top, we're building a system where everything's connected and searchable. And we think that's going to be pretty special. So that's my last bit. I'm sure some of you have commitments at the top of the hour, but as we put in the invite, I'm going to stick around as long as you all want and answer questions and can run past that if folks want. Thanks for paying attention. If you were on the webinar, we appreciate you coming if you're watching at home. That too. That's what I got. Okay, so I'm going to check out the Q and A panel. And so just a reminder, if you have questions that you haven't submitted, you can put that in the Q and A panel and then go ahead and upvote and I'm just going to answer them in order in terms of the ones that have the most votes. So a question that I see is how do you think the role of a KM person or team in an AEC firm changes in the shift from 1.0 to 2.0 to 3.0? That's really interesting. Good question, Jason. I'll start with where I ended near the end around the feedback. I'm just going to start with a thing like feedback. You have so much insight now into where knowledge gaps are within the organization. There was a little bit of, what would you call it, speculative knowledge building in the past that firms had to do. You had to assume what was important to work on. And now because you're getting that rich stream of feedback, you have a lot more insight and kind of evidence and conviction in where you should spend time, which is great for a couple reasons. It's great because you can be much more confident you're building the things that people actually need. But it's also great to go to a subject matter expert with evidence and say, like, you have the receipts, right? It's like, look, over the last three weeks, eight people have asked about this thing, like they're looking for help. And guess what? They probably came to you, right, and ask you for it. So you can really help. Like this helps. Again, I talked about that play like an orchestra idea. I think this evidence that you can bring of what people are looking for really helps fold those experts in and helps you build more targeted kind of knowledge. Knowledge content. But I'm going to. That's my quick answer to your question, but I'm going to kind of go away for a week and think about this and I'll probably put something on LinkedIn. So thanks for the great question. Roberto asks, is this search tool built around an existing LLM or is this proprietary? It's built around. So we're partners with Microsoft, so we're using Azure OpenAI. And if you really want to know, this is ChatGPT 4.1 mini. So as soon as maybe throw. We've got an AI search FAQ up in our help center and we can throw that in, which talks in more detail about the tools we're using. Thomas asks, being able to tap in all these resources is amazing. How does K, how do K clients, firms manage information that becomes out of date, like a PDF document or some. Yeah, so this is good. So. And then. Oh, sorry. And then is AI able to identify conflicting old versus superseding information? So there's a lot. I'm going to unpack that because there's multiple pieces to that question. So number one, from a very like, pragmatic perspective, a lot of firms run this. We have a. We have a process and maybe we can put that in the chat too, called Essential Content Review. And so it talks about kind of going through the different areas of content on your intranet with the content owners and whatever it is every six months, every year, just making sure stuff's up to date. You can also look at analytics and understand, you know, what's getting used and what's not getting used. So if you've got some content that's a couple years old that has like zero views in the last year, maybe it's time to archive that and pull it out of date. So those are a couple like process oriented pieces. Obviously people are going to give search feedback when something's out of date and so that will encourage somebody to either update it or pull it down depending on which is the appropriate action. And then on the third part of the question. So the algorithm for AI search prioritizes has some biases that we designed into it. So it does bias more recent content over older content. It will bias pages over like comments and social posts for example. So there's a variety of different ways and signals that we have to have a good guess of like what a higher quality resource to site is versus a lower quality resource in generally how long did it take from when a company decided to implement synthesis until it was normalized within the company? I mean that's a great question Brett, and maybe people want to put their answers in the chat. I think there's a first answer of implementation in general takes four or five months. But we kind of look at it as it probably takes a couple years to really get the kind of full breadth of the content. I mean this is a never ending project because your company continues to evolve and grow, but a critical mass of content and kind of all the process and the culture change, it could take a couple years. Some firms go faster, some go slower. It kind of depends a lot on culture in the existing quality of content with your organization. But I think you know, that's a, that's a range, I'd say one to two years before it feels like this is all the way kind of adopted within the company. Oh sorry Brett. Put a little follow up question in how long is the KA portion of the effort? Well that's a great question too. So we are there to help you run implementation but then we stay with you through the adoption process. So you get a client success manager from Knowledge Architecture who can work with you on strategy, on implementation, on driving change and adoption within the Org. So we are there as long as you have synthesis is the short answer to your question. So Jason asked, does Robocory have the self awareness to suggest when the searcher should actually make a real connection with a human expert versus using only the AI generated responses. It does not. I think that's a fair point. We haven't actually thought about that question. So that could be a kind of a thing it can evolve into over time. I mean one of the things that we do pass along to AI is some what the person's role is, what office they're in. So I think there's an opportunity over time to personalize answers more and more based on a variety of factors and this is a good idea of ways to think about using the fact they might be an emerging professional to maybe go a little more 101 level and a little more verbose and maybe if it's an expert asking maybe you can be just kind of assume some foundation. So great question, Jason. Nate asked a question about what sense does your system serve firm needs and procurement processes like proposal preparation. So that's a timely question. So people have. We didn't design synthesis to be a proposal preparation tool. It's a general purpose knowledge management tool. But we have seen a lot of adoption for synthesis, especially AI search for things like what we call more broadly a marketing research assistant. So yes, proposals, but also writing blog posts, you know, doing all these different activities that marketing communications do because it has such a wide range of content in it. So it is doing really well honestly in this. It's good at using unstructured information like past project sheets or boilerplate information or past proposals. One of the reasons that we're driving into this advanced employee search and advanced project search is just that so often what you have to do in a proposal is show how many projects like this have you done in the past who worked on those projects? When have these people worked together on a project? So that is driving a lot of where we're taking AI searches to be able to answer at a very advanced level questions about employees and projects and the clients and all that kind of stuff and how it's interconnected. So good question. We're going to talk a lot about that kind of work in the second half of the year as project search comes out and as we learn more and more from our clients that are using this, we're going to kind of like get them to share what they're learning in this community. So good stuff. SAPNA asked for firms using Microsoft products like Copilot. How do the two systems interact or do they not. They do not interact with. Currently a lot of our clients are figuring that out. Right. They're trying to give really good guidance to folks on when to use synthesis and when to use Copilot I wouldn't say that an integration down the road is off the table. It's just not been a top priority. What we do is, you know, I mentioned before that feedback. So we're seeing all of that user feedback from across all our clients come in and that is driving our roadmap. So that is why we did employee search after we released the basic tool. So the basic was 1.0, employee search was 1.1, project search is 1.2, and we think that the 1.3 is around the learning and management system. Right. Being able to search all that content really well. But I don't know what 1.4 is. So it could be that in our community it bubbles up that like really having an integration with Copilot or making that more seamless is important. It could be that we're doing something completely different. So we are not trying to plan out really any further than like three to six months in advance. And we're staying very close to our clients and understanding how they're using the platform. We do see, though broadly people using copilot and Synthesis very differently. Copilot is, I think, being used a lot of times as kind of like not like as a chatgpt equivalent. So like just like basically writing things, maybe some search. Although generally for our clients, the kind of things you're going to want to search besides email are in synthesis. So really Synthesis is the enterprise search platform, and Copilot is more at the individual productivity level. That's what we've seen so far. Frank asks if we have any plan to move to the next step, which is using the knowledge base to feed AI agents to do the work. I mentioned AI agents at the very end. The first kind of thing we're thinking about is, is around search agents. But I think that agents are the next big thing in AI and I think they're going to be really relevant to knowledge work. In my view, one of the foundations for doing agents is getting really good retrieval. Almost every agent use case you can imagine has to do really good retrieval. So we're putting a lot of our foundation, our foundational work now in getting really good retrieval. This is why you see like employee employees, like almost all the agents we can imagine or that we've heard about, ideas we've heard about from clients have involved project data to some degree. So we have to get really good at retrieving project data. The next step with agents is what are we going to do with that project data kind of on your behalf? Or what are you going to do with this firm wide knowledge on your behalf. So to say it maybe more succinctly, what we're doing now is foundational and kind of on our pathway to more agentic work. But we're starting with AI Search as an agent starting next week. Cool. So looks like I answered June's question, so I ran through them all. So thank you everybody for coming. I appreciate it. If you're still on the call and you haven't registered for K Connect, which is our annual knowledge management conference, here's a reminder to do so. We're going to be talking about how AI unlocks knowledge management in AEC firms. So thanks a lot and we'll be sending out this recording next week. Appreciate y' all.