INTRANET TOUR Refabricating DSX How Diamond Schmitt Is Evolving Its Knowledge Platform for AI Search

My name is Christopher Parsons. I'm the founder and CEO of Knowledge Architecture. You are joining us for an intranet tour where we share stories of how clients are using Synthesis Intranet and AI Search to help evolve and grow their AEC firms through knowledge and learning management. So we have three awesome guests today from Diamond Schmidt. We've got Jeanette Parker, who's the Director of Knowledge Management, we've got Wendy Gamal who's the manager of Knowledge Management, and we have Sam Horton, who's the Learning and Development manager. So Diamond Schmidt first launched their synthesis intranet dsx 7 years ago and we invited them to come talk to you today because we have been watching them kind of thoughtfully and strategically evolve DSX to capitalize on emerging technology like AI Search and our upcoming learning management system. And just the way and the process they've gone through, the thinking as well as the actual intranet itself we just thought was worth sharing with this community. They are a great representation of what we think the future of knowledge management and learning and development looks like, where knowledge and learning are integrated in terms of technology, in terms of teams, in terms of processes. So we're excited to have Jeanette, Wendy and Sam here and welcome. I'll let you all take it from here.

We're really excited to be here today to talk about all of this. We thought we'd talk a little bit about the firm first. We're a full service architectural firm with offices in New York, Vancouver and we currently have a staff over 300. We work in a variety of areas including performing arts, healthcare, higher education, research, libraries, recreation, community centers, transit and residential, so pretty much across the market. And I'm going to turn it over to Wendy for the first part of the presentation which is going to be launching the AI Search followed by how AI Search is shaping our content and preparing for the synthesis lms.

Great, thanks, Jeanette. We're going to start with a brief look at our intranet. We launched our synthesis intranet in 2018 and called it DSX, which is short for Diamond Schmidt Exchange. And then we launched the Synthesis 6 version of our intranet in February 2024. Our homepage probably looks familiar to many of you as we chose to utilize all of the research done by KA in setting up our site's architecture and the community navig. So we have communities for HR and office. We have a community for projects and practice, technology, business development and communications and learning, as well as a sandbox area for development. And you can see all of the menus as you mouse over you can see the breadth of the information that is available there. Our homepage also features our main news feed. We have some quick links to some highly utilized resources, and then our firm's Instagram feed as well. Now Jeanette's going to give you a much more thorough tour of our Internet later on in the presentation, and we wanted to start off today with talking about our journey to launching the AI search on dsx. So let's do a few demos of the AI search first. For those of you who haven't opted in yet or who are new to the AI search, you can still keyword search of course in the search bar, but now you can also ask natural language questions. For instance, how do I fill in an expense report? The AI search will generate a summary pulling content from pages and videos which are all cited within the text and displayed under the top results in the right hand column. Users can also search for resources on a topic such as Mass Timber, and get a detailed summary of courses from our learning program called DSU for Dynastrum University, reference materials from our library and DSX pages on the topic now that the employee directory is also included in the AI search, it mines information from our staff profiles as well so that users can be pointed to staff with expertise on the subject. In this search, our query is more directed towards the employee directory portion of the search, asking who our masked TIM member experts are. This also could have been a follow up question from the previous search for searches such as staff with specific certifications. Those of us who administer our synthesis platforms might think to go directly to the employee directory and filter for this type of information there, but general users don't always think that way, so being able to come up with the same information directly from the search bar is a big win. And then also we can let staff know that information such as expertise, certifications and licenses are now surfaced in the search, so that's a great way to encourage them to add content to their employee profiles. One of the most exciting aspects of the AI search has been its ability to surface information and knowledge from all of the video content that we have, whether it's design meetings or learning sessions. It makes your videos as searchable as documents or even better. As Chris was saying when we were discussing this presentation earlier, because the search takes you to the precise second in the video that's referenced. It is important to remember though, that not every search requires an AI summary. When someone's specifically looking for something like the letterhead, they don't have to wait for the summary they can just use the filters to go directly to the content that they need. Okay, now that we've looked at a few examples of the AI search, I'm going to take you through our process that we followed to launching the AI search on DSX. We generally opt in to KA's developments as soon as they're made available to us, opting into the video beta for captions and transcripts in July 2024 to prepare for the AI search public beta in December 2024 we created a roadmap to launching the AI search using information gleaned from other KA clients experiences such as EUA's AI search deep dive on their preparation. We noticed that some of EUA's lessons learned were no longer issues by the time we were testing. Because KA had already made adjustments to the search, we decided to use an iterative process with an ever expanding test group and continual content optimization. After the launch, we planned for ongoing maintenance, content review and staff training. We enabled the AI search for our first testing group in January 2025 and that was our DSX committee. We broadened testing in March of 2025 to a wider group comprised of a diverse cross section of the office, including members from Digital Practice, Finance, architectural staff, as well as staff from our New York and Vancouver studios. Input from our HR team was very important to us as well, but they didn't have the capacity to join the second testing group at that time. We met with the test group and explained to them what was included in the search. We showed them how to rank and give feedback. And note that at the time of this second testing group, the AI search did not include the employee directory and we communicated this to them, but they still gave feedback on content that they felt was missing from the AI summaries that would have come from the employee directory. So it's important to make clear what is searchable and what isn't, to manage expectations, and also to remember that general users are not necessarily going to remember what's searchable and what's not. We broadened the search testing group again in May and included representatives now from our HR group and our AI committee. Now I'm just going to talk a little bit about what we tested. For the initial testing done by our DSX committee, we established a priority list to identify what we would test. We started with the search term list provided by Knowledge Architecture and tested using a variety of keyword searches and natural language queries, and we split that list up between our DSX committee members. Next we tested DSX Top searches based on the final search analytics that we had from Synthesis five. And then finally we tested searches based on top page views and common questions that we receive in our various roles. But for the broader testing groups, we decided to let them form their own queries and keyword searches based on their own areas of expertise and the questions that they frequently received. The early days of searching yielded some interesting results from time to time. In this query asking for the members of the Finance Committee, the name at the bottom of the list Jennifer o' Connor that has the citation beside it was not in the page that was cited. In fact, we didn't know who Jennifer o' Connor was. This name came from a video from an external presenter. We were also confused why the Social Event page was the top list in the results as well. Luckily we could provide that feedback. So the search Optimization panel now includes a feedback tab and this allowed us to rank our search results and record feedback as we went. The ratings and feedback allowed us to track our results. What we needed to optimize update content that was missing from DSX but also helping to fine tune the way the search works. We received emails with direct feedback from KA throughout the process based on the feedback that we recorded in Search Optimization. In this instance, Greg had reached out to ask if the issue was that the information in the AI summary was not on DSX or if the information we were making a note to add to DSX in the future. I also wanted to give a shout out and a huge thanks to Adam from our BIM group who did a lot of thorough testing of our technical content. As many of our DSX committee members are not architectural staff, we wouldn't have been able to evaluate the validity of the search results and summaries and needed to lean on their expertise. We were pleased to see users in our test groups asking follow up questions to refine their searches. We were training the AI, but the AI was also training us to help form queries to get the best results. Results. The first week of June we evaluated all of the results, ratings and feedback. As you can see, a great majority of the ratings were in the four and five star categories and keep in mind that a portion of the lower ratings would have been for searches for content that we simply didn't have available on dsx. Jeannette had met with HR to do test searches with them and then I met with HR again to go through dated HR or policy posts to make sure there weren't any concerns. Since we had done a huge cleanup of our wiki page content prior to launching Synthesis 6 in 2024, we were mainly concerned with dated content that might be pulled into the AI search summaries from posts. So we decided to launch June 9 and communicate the new feature to staff with announcements at meetings, posts on DSX and ongoing training. I announced the AI search launch in person at our weekly design meeting in Toronto, which is available virtually to our New York and Vancouver studios as well with the following short presentation. I explained ka's methodology, what was currently included in the search posts, pages, videos, the employee directory, and what wasn't the project directory. I showed them what the AI search would look like on dsx. I demonstrated that you could now search with natural language queries and showed them how the search results would appear with a summary with filters and the sources of information. I showed them how all of the information is cited and it links directly to the source information. I showed them how to ask follow up questions to fine tune their results. I also explained how to rate and provide feedback and explained to them that it would not only help us improve dsx, but it would also help KA improve the AI search. I also demonstrated how the AI search surfaces all of the knowledge included in our video content, which is consistently one of its most impressive features. I showed that the video transcriptions are searched, allowing the citation to take them directly to the reference spot in the video for just in time learning making our videos as searchable as documents. And then finally we revealed the launch date which was the following week. Our DSX committee member Matthew in Vancouver also announced the launch in person at their weekly meeting. We then followed up with a post on DSX and included all of that information from the design meeting presentation for staff who weren't able to attend either in person or virtually. And finally we have some lessons learned that we wanted to share. From our optimization of the AI search, we learned that we should use post to promote page content instead of putting all of the information in the post to avoid having to review posts at a later date. We're working on formalizing our content reviews and updating page content on a scheduled basis. And we'll continue working with HR whenever there is a change in a program, such as a benefit or a policy that gets announced in a post to remove previous posts on the topic if necessary. For instance, the HR group posts on DSX to communicate something like a cost change in benefits, and understandably they want their post to contain useful information and not just be a link to a page. So when a change happens we'll look at the previous post on the topic and perhaps delete it if necessary so that that older dated content isn't pulled into an AI search summary. We created the hashtag KMreview to tag content like this, as well as any other dated or temporal content we might want to consider removing at a later date. We are going to continue using Best Bets to prioritize content and direct users. Best bets indicate content that is vetted and increases users confidence. We also learned that we can use the AI search as a tool to encourage people to fill in profile information, since that information is now surfaced in the AI search and we'll use that same logic to hopefully get people to give us more project information as well. We learned that structured content works very well for search summaries, e.g. wiki pages with clear headers and sections or videos with chapters. And again that the video transcripts and chapters make our videos able to be comprehensively searched, making all of that valuable content more accessible. The Feedback panel enables continual improvement to the content on dsx, providing a direct, immediate link for users to provide feedback on content or missing content. We'll continue with our series of did you know posts on DSX to train users and communicate features of the search and changes, and we're introducing new employees to the AI search as part of our onboarding. Now I'm going to stop sharing and Jeanette will give you a deeper tour of our intranet and how the AI search is searching shaping our content going forward.

Thanks fundy for our tour. It's not going to be the most in depth tour because of time constraints, but I thought I'd give you a little tour of how our DSX is structured and some of the types of content within. So but first, this is our DSX Core committee made up of members of our Knowledge Management Group, Learning and Development and Architectural group. And this group has Everyone on the committee has a passion for learning and sharing knowledge and we meet bi weekly to review metrics, share ideas, set goals and ensure that we're providing relevant content and continuously improve the experience for our staff and onto the tour. Our home community is where we have our information for incoming staff under welcome. Also useful for staff who've been around for a while. Our about us information with our communities, groups and departments, the directories and then information about events and our community outreach are all on this page. And as Wendy mentioned, we have the quick links on the side for our frequently used content like our event schedule for all of our learning and development events in the office and our Instagram feed. Our HR and Office section has all of Your typical HR content, employee handbooks, information about reviews and con ed office resources. So anything to do with our general office, where to get letters and templates, how to ship a package, booking meetings and then everyone's favorite thing to do, booking holidays. So I think that's one of our most checked out spots. And Finance also has information on expense reimbursement and invoice approvals and that sort of like our third community I think has our most developed content is Projects in practice. The project delivery section is a phase based section for our project managers who are looking for information on running the projects. And then our reference area which is populated by the Knowledge management group has hopefully just about every single piece of reference material that you could want. And then our design resources area is all staff created content based on lessons learned, information we've gained and expertise from all of our projects. This is a great resource for the staff and I also want to give a shout out to our digital practice group for their excellent content. They have curated an amazing amount of content documenting best practices, workflows, softwares and other guides that enable the teams to spend more time designing and less time on figuring out new workflows for each project and making their lives easier. So and then our third, well fourth really is our technology section where we have information just about every piece of software we've got it best practices, our expanding AI section and also how to get help when there's those times when your technology just does not work as you intended. So our business development community has our brand guidelines, communication plans, all the logos, ingredients that you could ever want and links to open asset and templates for presentations and reports. And our the last community we're going to talk about because our sandbox is just laboratory experiment is our learning community. Sam is going to expand on that during her presentation, but this area has all of our course content from the past DSU sessions. That which is Diamond Schmidt University that we've done and it is 100% created in house and is full of our staff's expertise and ties in really well with the rest of our content. So then we wanted to know how AI search is affecting how we're structuring our content. And we have a few things because we have a much deeper access to the content than before with regular search. We're reviewing our content review process and luckily we had done a huge content review process when we moved from synthesis 5 to 6, but really getting a more fulsome and more frequent content review with the content owners to make sure that we have the best content available. And these reviews are going to start to include the search feedback in the review process so that the content owners will understand what people are looking for and then giving them some guidance on how they can optimize their content for the AI search. And as Bundy was talking about posts with old data in them and information. So we're putting together a periodic content review process for posts that includes helping us identify the temporal posts and other posts for deletion with the KM review tag and then creating a transparent process and criteria around guidelines for removing posts so that we don't have anyone that gets upset if a post has been removed. And then also encouraging staff to link to information rather than embedding the information in the post because we have had issues where old PDFs have popped up out of nowhere in the test searching that we were doing. So the next part is on structuring our content. So again, we have done a fairly thorough job and we upgraded our platform. But reviewing the content, how it's structured to optimize it with the AI search and making it more uniform and also reviewing the language because sometimes people have written things that are a little bit wordy and could be a little bit more concise. Another part is we have a number of PDFs on the site that could be translated to wiki pages and we've been working with the content owners to do that because having the content as guides is easier to manage, better surfacing the data and has the ability to include more dynamic content such as automated post blocks and automated project blocks inserted in the content. So we have an example where that happened. We have a drawing manual in the office that was a PDF and our digital practice group worked with the creators of the digital digital drawing manual to really bring it in as a guide. And we think that it's excellent, easier to update. And also now we can give the group more insight into if they're how their material is being used and what the more popular pages are for working on with the drawings. Another part of our content is restructuring some of the pages in the Design Resources section. These pages have a lot of good data on them. This Mass timber is an example of one. It's in a work in progress. We recently through Diamond Schmidt University, completed a series of talks on mass Timber and we're incorporating that information into these pages. But we're also taking this opportunity to work with the subject matter experts to come up with an expanded layout and guide to include all of the information from and best practices that they've been coming up with as part of this series and leveraging that knowledge base. It's. This is a work in progress. This is kind of the thought that we're actually meeting with them next week to talk about whether this type of format would be best. And so hopefully it will be. We're kind of excited about it and hopefully we'll be applying this to other materials and systems as we go. And that leads into a coordination of efforts. Our Project Delivery committee has been really impressed with the AI search and the work that we've been doing on bringing together the content with the mass timber through the DSU sessions. So we're working on that when there's sessions on series on glass, masonry, brick, different building systems, that when those series are happening as part of DSU that the technical subcommittees that are part of the Project Delivery committee will feed into that and help create a critical mass of knowledge to create these really well developed information guides that we can use through the AI search. Another experiment that we're doing is in the more bite sized content. In this particular example, we've had a film student with us this summer who's been going to the different sites reasonably locally with our architects and filming content on site. So this particular one was about steel studs and the architect explained in two and a half minutes a fairly detailed account of what you should be looking for, where you can go wrong and giving a mini lesson. So now that we've put this bite sized stuff in a post, whenever people are looking for information about metal studs, a lot of that knowledge will come up in the search summary. So these posts have been very popular and we're hoping to continue them. And another part with our kind of bite sized posts is our staff are really great at updating us on projects when they go to site and oftentimes there's a lot of really good information that could in the past have sort of stayed a little bit buried and not be incorporated into the searches as much. But now with the video surfacing and the AI search that when you do a search on a particular topic it'll bring it up in the, in the search summaries and we think that's a really valuable tool. So I'm going to just switch to our live site here for a moment, so bear with me because our staff sometimes get extraordinarily technical in these posts and they do put a lot of effort in and that information is so valuable, this one could not screenshot very well. So, and it's a particularly good example so one of our architects, well, actually the whole team went out to look at terracotta facade mock up at the Art Gallery Ontario where we're doing an addition to the Art gallery. And as you can see, he's put a lot of information in here with images, he's done videos of the mockup. And then closer down to the bottom here, he's put a bunch of very thoughtful observations about the mock up in the process and then even included the write up that was presented to the city as far as the materiality and the choice of the terracotta for the building. So now this information is dynamically being included in these search summaries and we think it's absolutely fantastic and are encouraging the people to do even more stuff like this. On a final note, with shaping our content, this was a bit of a side bonus, but I thought sourcing new content. One of our architects was super impressed with how the AI search worked. He actually wanted to apply it in a whole bunch of different scenarios, especially to do with healthcare. And he has actually offered to curate an entire healthcare section, enhancing all of the information we have. Now, most of our healthcare resources are just to do with standards and information like that. And he's offered to do a whole lot more because he's so impressed. And this is just a shot of a preliminary draft with a talking point that we're going to be meeting with him in a few weeks to see what can happen. So I think more people are going to be coming forward with that as well once they see the power of the EA search. So that's it for me. I'll stop sharing my screen and it's over to Sam to talk about learning and development.

Awesome. Thank you, Jeanette. So hi everyone, my name is Sam Horton. I'm the learning and Development Manager at Diamond Schmidt. And today I'm going to talk a bit about how we are preparing for our Synthesis LMS rollout and how we're thinking of it not just as a tech update, but as a way to better support how our team actually learns. So I'm going to walk through a bit about our approach, what we've learned and what we're excited to build. So by now you're all familiar with our version of Synthesis dsx. So I would like to introduce you to dsu, also known as Diamond Schmidt University. So we launched in 2018 and we started small with just a couple of guest speakers and since then it has grown into a full fledged learning and development program. So we strongly value in person, social learning at our firm and, and that's why we prioritize formats like labs, charrettes, design competitions, tours, workshops. We run the gamut. And so each time we host an in person learning session, we capture and stream it over Zoom and then we take that recording and we create an on demand course on dsx. So we follow the blue line. If you guys are familiar with this. Before synthesis 6, we used to upload our recordings to Vimeo and then share those links in DSX posts. And it kind of worked. We chapter the videos and add resources like you're seeing here. But the courses were hard to find and there wasn't really a central library or structure for us to follow. So we introduced a learning management system, Absorb lms, which some of you might be familiar with. We spent months building every course from scratch, building the platform more or less from scratch. Scorm files, thumbnails, and it looked great, it worked great. And then nobody came to our party. It was too many clicks. People didn't want to leave dsx, log into another system, learn a new interface. So after two years we sadly had to shut it down. But all was not lost because then synthesis 6 came along. So for the first time we could consolidate everything, upload directly into dsx, give each course its own page and create a proper catalog for our content. And with the AI powered search that Wendy and Jeanette have talked about, every single video transcript is now discoverable. And that's something I'm still wrapping my head around and I really want to pause and emphasize that. So our program has been around for seven years. So that's seven years of specialized knowledge and learning material that otherwise might have been buried in posts or forgotten. And it's now accessible to every staff member with a couple of words into the search bar. And that is amazing. And that's the power of AI in learning. But it didn't happen overnight, at least not for us. So over the last six months or so we have been migrating over 200 existing recordings from Vimeo, uploading them to DSX, re chaptering them and giving each one a course page. So we've been doing this while we've been running our live program as usual and have just been doing it as an incremental rollout. And this has been made possible thanks to Rowan and Avery from our L and D team who are making this all happen. So big shout out to them. Now I want to move on to a few elements of the LMS that we are excited about. The first being on demand content. So that refers to courses that are specifically made for the lms, not just recordings of live sessions. So as we all know, watching a recording can feel a little bit impersonal because it wasn't made for you. You're sort of catching up on something that's happened where on demand content is designed to be clear, focused and easy to access when you need it. And something that we really want to lean into is that it doesn't replace social learning, but it lays the foundation for it. Because if you can cover the basics on your own, then you can free up time to go deeper with your team. So for example, you might watch a four minute short like this one on how Glass Curtain wall is installed and then you can join a lab session to apply that learning. That might look like detailing it in revit, might look like testing mullion spacing with different models or tons of other hands on applications. And longer term, we would actually love to see this learning sort of downloaded into teams where they start using these shorts as part of their project work. So maybe that looks like watching a short relevant video together and then seeing how they can apply that in their model in their drawing set. So there's that sort of in person mentorship happening as well. So that's the dream. We're also excited about learning paths, so we've piloted a few so far. BIM 360 and Jeanette mentioned our Mass Timber series as well. But most of our content is standalone, so the LMS gives us a way to group our learning more intentionally. We're considering role interest sector, everything's on the table and as Jeanette mentioned, all of our DSU content is made in house. But we are excited to expand our library with external courses as well. In particular, we're interested in software upskilling for applications like Power BI and Excel soft skills training. And eventually it would be great to continue our consolidation trend and integrate our mandatory training like whmis and cybersecurity that we are currently outsourcing. And finally, we're looking forward to more robust metrics looking at which courses are getting the most traction, how usage varies by department or role or even studio, and how people are discovering content as well. Are they moving through the course catalog? Are they exclusively using the search bar? How are they sort of finding what they need? So we're going to be looking at all of that during monthly, monthly reviews after we launch to help us understand what's working and what we're going to build next. So if you are just starting off on your LMS journey, I would like to share three things that we've learned so far. So with all this new functionality, it might feel overwhelming. I think it did to me at the beginning. But you don't need to start from scratch. If you have a learning, new management or learning and development program, chances are you've already got great content. So just start with what you have and use the transition as a moment to review your content. So some of the sessions you'll retire, some you'll refresh, and some you can build into, you know, clearer learning paths and then finally communicate early and often about both the AI powered search and the lms. Because it's not just another tech rollout, this will be a shift in how all of us learn. So the goal is to spend less time searching for information and more time, you know, learning together, collaborating together, solving project challenges and pushing our design forward. So on behalf of Wendy, Jeanette and myself, thank you. And I will throw it back to the KA team.

Awesome. Thank you, all three of you. That was really great. It gave us kind of three different lenses into how you're doing knowledge and learning. And I want to just start with a question around prioritization because there's, you know, Diamond Schmidt's a very dynamic firm with a lot going on. And so both on the kind of the knowledge side with the intranet, maybe more for Jeanette and Wendy and then for Sam on the learning side. How do you decide which content you build, in which order? How do you decide from a learning perspective what learning that you, that you go pursue?

I could talk for a little bit. So, I mean, one of the things that we've always done is very happily taken any content that anyone gives us to try and build on it. I mean, if someone wants to offer to make us a page, we're super happy to take that. I think what I'm trying to focus more on as a priority is getting building on some of the content that we have, because we have a lot of, we do have a lot of standalone content, you know, some details about a certain thing and that. And, you know, someone's done a white paper on heritage building, but it's a little bit disconnected at times. I mean, it's all extremely good content. So that's one of the reasons we've been working with our PDC group to get some prioritization on content there on our technical content, because that's what we do really want to grow because that's the most useful stuff to our, our staff. And we've, we've recently Doing that collaboration with DSU that really getting people to work together to get that critical mass together instead of having these little bits and pieces coming in. So I think really trying to get that technical content so that people don't have to be repeating research and starting from scratch. Every time.

You say disconnect it, do you mean that it's kind of like advanced kind of content that's kind of floating but like context that explains what it is, isn't there?

It's more that, you know, someone put a lot of effort into doing a reference detail on glass, but that's our only glass content or you know, so that it's, you know, and that, you know, it would really benefit from having other content in with it. So just getting people to really get that critical mass together because we've got lots of content creators and we're really trying to prioritize getting them organized and behind topics to get, you know, really comprehensive guide to whatever it is that we're working on. So. But yeah, it's, I think that's where the prioritization comes from. Sam, I was just going to say.

Maybe just to add to that a little bit. A lot of our committees will have initiatives and a lot of those will be end up being presented to the staff through dsx. So sometimes there'll be, you know, a detail on one thing or a manual on another thing, but they're not necessarily connected in any way. But as I'm hearing this, I'm thinking the great thing about the AI search is if you search on that topic now, you are going to get those, those all pulled together. So that's helpful as well.

I would say trying to align with our strategic goals when, when possible. I mean mass timber as you can, as you can tell, has been quite a focus of ours over the last few months. So we're making sure that, you know, those courses are top of mind, posted regularly built out well referencing the manual. And even when we are, we can sort of cross collaborate where when we're showing the functionality of the AI search, for example, we reference. Okay, where can I find, you know, mass timber learning content? So we're sort of working hand in hand to, to reinforce that.

So you're using, you're kind of, you have to pick something to demo AI search to the staff. So you're picking something strategic. So you're also reinforcing mass timber even through your demos is what you're saying.

Exactly. It's a hot topic that people know is being focused on in different sectors of the Office. So might as well use it as a, as a jumping off point to get people excited about the AI search.

Amazing. I don't know exactly who this for, so I'm going to throw it out and whoever wants to jump on it can. There were a couple questions around frequency of content reviews. So how often should content be reviewed and by whom? Like maybe we can just talk about content review process stuff.

I can go if you want or Wendy, you can go. I'm happy the other way. I mean I think at a very, at a minimum we should be reviewing our content once a year. That's the minimum. You know there are, there are some sections where maybe more frequent reviews may be required or sometimes less because I know our digital practice group is they're constantly reviewing their own content and adding them to it. But I think at a minimum once a year, ideally if there was time because we also know people are busy for some groups maybe twice a year if the, if it's the content that may be changing frequently or becoming outdated quickly.

Yeah, I just wanted to add to that as well. I know I talked about things like HR content when we were talking about going forward, the AI search and that's something that's just going to always be updated when it's needed. So there's content that's fairly static and then there's content that may be fairly static but then every once in a while there'll be a big change like oh, our benefits packages have changed so we need to go and look at those posts to make sure we're not pulling in dated information. The pages need to be updated. So that's not something that needs to be necessarily like a quarterly review but just, you know, when there's a change we need to know about it. We need to update everything. So I think some things, a quarterly six month, a yearly review would work and some things just have to be more ad hoc.

This is sort of adjacent, but I do, I'm keen to add it in as part of that transition from sort of longer form in person, you know, lecture style learning that you're then recording. I think another one of the benefits of sort of breaking it down into those bite sized on demand pieces is that as you go along, if something in that you know, is out of date, maybe it's a particular element of that software that's changed, changed. It's much easier to go in and swap out that four minute piece for the 2025 version rather than revisit. Okay, do we have to re record this entire session? From 2021 to make a few tweaks. So I think there's definitely some opportunities there in terms of content review and keeping things up to date.

Yeah, it's kind of like the long, the longer a video is, it's an asset, but it also can become a little bit of a liability if you kind of like, look forward.

Exactly.

Yeah.

Yeah.

We had a question from Carrie at bwbr. So you had mentioned these kind of project update posts. Like, is there any plan to move some of those to pages or guides if they're not kind of just like a project tour, but like, here's like how we designed this thing. Like, do you think that those might become more like pages or guide to help them feel more evergreen versus having to maintain posts going backwards? You know where I'm going with that?

I ponder it all the time, especially with some of our, I guess, more prolific writers, that it's almost like they've written a book, which is great. But, you know, with the advent of the AI search, I'm not as certain that we necessarily want to do that. It may depend that, you know, when we're looking at a topic, maybe we're polling the posts and in some cases, maybe a particular post may be a good idea to turn into a page to fold into that information. But I think it's going to start coming up as we're adding more of these guides together because, you know, it's occasionally standalone information, but maybe if it's more tied in with a topic that it could potentially get. Get tied in. So it's. It's one of those things that I actually think about quite frequently, and I just haven't come to a firm decision on it yet. So I think we'll see how that goes.

Yeah, we had some questions about content ownership too. Like, how did you establish who your content owners are? How does that change over time? And I think that could apply even to the learning piece too, in terms of maybe the person was the instructor, but no, they're no longer at the firm. So how do you stay on top of connecting people to knowledge and topics?

I think we did a lot of that kind of establishing ownership. When we were moving from synthesis five to six, we really looked at who the owner of the content was, who is the creator, who would be the creator moving forward if it was current, if we didn't have an owner, Sometimes we just parked that content and moved on. If there wasn't anyone who could confirm to us that it was up to date. And. And we do do all of the page, the page creation ourselves for the most part, except for our digital practice group. So everything does kind of get vetted at the time as well. So we kind of tried to assign who we think the content owner is. I think sometimes we would go to them with the content and they'd be like, oh, that's not my content. I'm not responsible for this. But generally, you know, who's in charge of which content just by the groupings on the site as well.

Anything from the learning side, Sam? Or is it the same kind of process or is there a nuance that's different?

I would say it's pretty similar. And I mean, we do the content, the learning content review as part of our general content review. I also, I think this is just a, you know, nuance of how Diamond Schmidt is structured. We were very committee based. So often people are already sort of grouped by, you know, specialty, you know, the sustainability committee. There's someone looking at detailing. You know, there's sort of a subset for everything. So there's typically not just one person to go to or if they're away or if they've left, then you know that there's an ecosystem of people that you can, that you can draw from.

Curious about the short videos, the shorts that you showed and some of the kind of the prototypes of those. What's the lift like on that? You know, like, are you fully scripting them in advance? Are they kind of ad hoc? And people ablumg on site like to get to three to four minutes? Sometimes it's harder to make a short video than to make a long video. And so I'm kind of curious what that's looking like.

Definitely they are brand new. We started posting those this week. Oh, wow.

Okay, great.

So it's very, I would say it's loose. We've sort of tried to test a few different approaches and I think what has emerged is that people actually prefer when it's not too slick. That sort of FaceTime vibe does suit that shorter content. So how we're currently doing ours, we've got two GoPros. One GoPro, you stick it on the helmet of the person touring you around the site. And the other is our film student, Avery. And he sort of in Adobe Premiere, flashes back and forth between perspectives, depending on what's being shared at the time. And we try and keep them five minutes or less, ideally less. And the lift really isn't that bad. It's just switching between two perspectives. So you need someone who's relatively handy in Adobe Premiere. But after that it's very reasonable, I would say just as reasonable as trying to edit sort of a longer form video.

Right on. We've got kind of a question going completely in the other direction. Somebody in the, in the chat had noticed an approved AI services on one of your slides and they were curious what that is and how that list is determined.

Okay, I can speak to that because I'm on our AI committee. So the approved services. Because in our, we're experimenting with all sorts of different types of AI and we don't wanted to get ahead of people just randomly going off and throwing entire projects worth of stuff into the free version of ChatGPT because not such a great thing. So the committee got together, we've gone through a number of different AI tools where we've vetted things like the security, having paid accounts for them, what, you know, in I think some cases, you know, what you can showing what you can and cannot do with maybe a free version. Like if you're going to use the free version of ChatGPT, you have to pull a whole bunch of stuff out and get identifying information in. So it's a dynamic list where we are really trying to look at which services, which tools are the best and how we're going to use them. But we wanted to give staff some guidance instead of them just picking and choosing perhaps some of these free things that may or may not work very well with our end goals and letting our information leak out there in unauthorized ways. So it's, it's an ongoing process and. But there's a few different ones that we've identified as these are good to use and then we're constantly taking suggestions and evaluating different tools.

Nice. I think we've got time for one more question and we're back to content review again. So this seems to be like, obviously we're going to need to do something about content review and a deeper thing in the future. Although Susan did put something in the chat about our essential content maintenance plan that I recommend that people check out that resource. It's really good. But the question from Carrie is around how are you tracking all of this stuff, like keeping track of who the page or content owners are when you've last reviewed something, what's flagged for update, like what are the mechanics look like?

I have an Excel spreadsheet that's hard to maintain. So I mean some, some tools would be, would be fantastic. I think being able to, you know, have the reasonably because sometimes, you know, you're almost reviewing something on the fly with somebody. So you know that. Oh well, we just updated together this page. You know, it doesn't need to be reviewed for a while, but it, yeah, something would be excellent. But yeah, right now it's an Excel spreadsheet, which is. And then, you know, you have to think, oh geez, I added five pages. And so it's a little cumbersome.

So I have to say in the search optimization, having that feedback tab is amazing because if somebody, you know, comments and said this content is out of date or we're missing this content, it's a big trigger to update that content. And plus it's all tracked there for you and you can check off when it's reviewed. So I think that that feature just of the AI search will be really helpful in maintaining fresh content as well.

It's amazing. So thank you all. We covered a lot of ground in a short amount of time. So thank you Jeanette, Wendy and Sam for sharing your work with us. I think Susan put it in the chat earlier. The recording of this tour will be put on our website in the online KA community now next week. And we look forward to seeing you in future events in the K community, including K Connect in a few weeks. So thanks everyone for attending. And again, one more thank you to Diamond Schmidt.