Introducing Advanced Employee Search

Hi, this is Chris from Knowledge Architecture. I'm going to introduce you to Advanced Employee Search, our latest synthesis AI Search Feature Synthesis already contains rich information about employees brought in via our integrations with AEC software platforms like Deltec, uninet, Open Asset, AEC360 and Azure Active Directory, including employee profiles and directories which contain structured data like departments, disciplines and offices, licenses and skills and experience. We have curated employee blocks on intranet pages which identify the go to person for a given topic as well as employee mentions in intranet content like video transcripts, documents and posts so the opportunity was clear. Combine that structured and unstructured data with advanced AI capabilities to help you find the right people or with the right expertise at the right time. Let's take a look at Advanced Employee Search in action. Who are our registered architects in California? AI Search is categorizing the query to determine which tools to use to answer the query and you can see that it generates an AI Search summary and there are six registered architects in California. And if I hover over the citation, I can see that it's citing the employee directory as the source. Here are search results for who are our pes in Washington, who's on the marketing team, who's on the sustainability committee, and who works remotely, all of which AI Search cites the employee Directory as the source of truth for the answers. Now here's the search result for who can help me plan a flexible Wet lab, which suggests Ann Johnson as a knowledgeable resource. And if I hover on the citation, I'll see that Wet Labs page we saw earlier as the source of truth because of this employee block on the page. Here is the search result for who is our Mass Timber expert. Synthesis AI Search suggests Julie Lopez citing her employee profile, a video she created on Mass Timber Basics in a recent post about Mass Timber Basics as sources of truth for this answer, here's the search result for who is our Patient Centered Care expert? Synthesis AI Search suggests Carla Lewis citing her employee profile, her resume from the Resume library, and the Healthcare page from the Marketing Boilerplate guide. And finally, here's the search result for who can I ask about reclaimed structural steel synthesis? AI Search suggests Ada Lovelace citing her employee profile as well as a recent post about a project tour. In addition, Synthesis AI Search suggested the Sustainability Guide page for Building Structure as well as contacting Julie Lopez based on her role as Sustainability Director as well as regular contributions on sustainable materials and embodied carbon. So you can see that we've made AI Search much more powerful and that advanced employee Search will unlock the ability for your employees to connect to the right person in the flow of their work. Oh, and one more thing. If you do a general knowledge search like Revit Container Files Synthesis will also recommend go to people for that topic when the data is available. As you can see here, AI Search is suggesting to reach out to Brianne Walton for further questions or details. The Digital Design Manager and I can see that if I navigate to the Revit Container Files page. Brianne has been suggested because she's listed as a contact at the bottom of the page. Cool. Okay, consider yourself warned. I'm about to geek out a little bit, so if you leave at this point, I won't be offended. Okay, still with me? Here we go. I want to spend a minute on something called Retrieval augmented generation, or rag, if you're familiar with the term. And just to clarify, for folks who may not be, we're not training a model on all your firm's knowledge and asking it questions. Instead, we do a search first to gather the best possible information. Then we pass that information to an AI model called a Large Language model, or LLM to generate the search summary. That's RAG retrieval, because we search first, and that search augments the generation process of the search summaries. The retrieval step is also what allows us to create citations. The next evolution of that and what we use to build advanced employee search is something called agentic rag, or what some people are calling search agents. So what's an agent? A search agent is a system designed to act on your behalf, to work towards your goals, not just return search results. That's where the word agent comes from. It has agency. It doesn't just look things up, it thinks, it acts and adapts. To do that, an agent needs three essential capabilities. First, it needs reasoning. This means it can interpret the intent behind your question, not just what you typed, but what you actually meant. It can figure out what kind of question you're asking, or even whether it's actually answered it, or if it needs to keep trying. Second, it needs tools. These are things the agent can do on your behalf. In the case of employee search, that might include filtering the employee directory, searching the Internet, or combining both strategies to find the right person. And third, it needs access to knowledge. In this case, the content on your intranet and in your directories. With reasoning tools and knowledge, a search agent can do more than just retrieve results. It can help you achieve your goal. So let's take a closer look at how that works. Inside synthesis, a search query comes in, say who can help me with the drone survey? We first send that to the large language model and ask what kind of search is this? Is it a general knowledge query or is it an employee search? It uses reasoning to decide if it is likely a general knowledge query AI search will run a standard rag search like we saw earlier. Who can help me with a drone survey? Appears to be an employee query. So synthesis tells the agent, great, you've got these tools at your disposal. Search by employee name search, role or title search expertise. Just do a directory filter. In the case of this query, who can help me with a drone survey? It's expertise. So it will search both employee profiles and intranet content in order to generate the best answer. If the results aren't strong enough, it loops and tries another approach. It keeps going until it finds an answer it's confident in. So you're giving the AI tools knowledge and a goal and it uses its reasoning to help navigate you to the right answer. This is how Advanced Employee Search works and this is why it's called Agentic Rag. Now, one important caveat While Synthesis Advanced Employee Search can answer a wide range of questions using employee directories and profiles, employee blocks on pages and text in pages, videos, documents and posts, there's one important source of truth we haven't yet the Project directory. This means if you ask questions like which employees have worked on higher education projects or who has worked on MOB projects with Sutter Health or who's worked on K12 projects with Vandelay engineers, Synthesis AI can't answer those yet because it hasn't been taught how to map employees to projects and clients using structured project data. The good news? That's exactly what we're working on next projects will be the next major capability for Synthesis AI search. As I've diagrammed here, you can see that projects have been added to the query tree that Synthesis AI Search uses to work through searches, and that we're adding project specific tools such as search by project name search, employee project relationships, search, project experience, and directory filters. Here are a couple screenshots from the lab of what we're calling Advanced Project Search, which is in development now. What higher ed projects have we built in California? So Synthesis Search the Project directory, and that data is sourced from our integrations with systems like Deltec, Uninet, AEC360 and Open Asset. Or just to show off a little bit more, what higher ed projects have we built in Massachusetts which were over 200,000 square feet? Or please list our Leeds Silver or higher projects and you'll notice in the search summary that AI Search figured out to include silver, gold, and platinum projects. Nice. And finally, what projects have won AIA awards? So pretty cool, right? As I said, this is in development and will be heading into beta once it's ready. Advanced Employee Search will be available in public beta on June 6th. If you're already part of the AI Search beta, it'll be turned on automatically at the end of the day, and we'd love to hear what you think. If you're new to synthesis, you can learn more@knowledge-architecture.com we'd be happy to show you what's possible. Happy Searching.