What if you had an internal search engine that could access and easily retrieve all your collected data and information? A tool that could understand unique AEC industry terminology, and help us find what we need, not just what we asked for? The future is now with artificial intelligence (AI) powered search, the topic of this edition of The Friedman File.
The Institutional Knowledge Quandary
Our industry thrives on institutional knowledge. However, our valuable experiences, project summaries, technical details and lessons learned are scattered arbitrarily across our digital ecosystem. Information lies in different software platforms, with inconsistent, convoluted naming conventions. Traditional keyword searches can’t decipher this puzzle, leading to inefficiency and lost opportunities.
AI-powered search can help solve this challenge. By integrating AI into intranet search functionality, firms can unlock the collective intelligence of their organizations, more accurately retrieving contextually relevant information. AI search also aggregates data from the various platforms that AEC firms use (e.g., Deltek, Unanet, Newforma).
With AI search, AEC professionals can capture massive amounts of institutionalized knowledge from a single, accessible location and receive it in an organized, understandable way. Firms can more easily share information across disciplines and locations, and ensure that hard-earned lessons and educational experiences aren’t buried in network rabbit holes where no one goes.
The Evolution of Synthesis AI Search
Synthesis AI Search is currently leading this transformation. Developed by Knowledge Architecture (KA) and included in its intranet offering, AI Search represents a significant shift in how AEC firms interact with their knowledge.
Synthesis has evolved from its origins as a SharePoint-based intranet solution to the AI-enhanced version currently in public beta testing. KA founder Christopher Parsons says they rebuilt their product from the ground up after recognizing that intranet search tools struggled to retrieve institutional knowledge efficiently.
“We’re moving from keyword search to a natural language search, so people can ask the question they really want to ask,” says Parsons, who founded KA in 2009. “This requires that AI understand the intent of the question and fill in what might be missing. This gives us the best chance of retrieving what we’re looking for.”
This breakthrough was driven by retrieval-augmented generation (RAG) — an AI approach that retrieves live search results, feeds them into a language model, and generates summarized responses. “RAG prevents hallucinations because it is grounded in real data. It doesn’t editorialize because it can only draw from the data it is given,” says Parsons.
Synthesis AI speaks the language of AEC — so it knows that “Revit” isn’t “rabbit” — by including an advanced AEC-specific video transcription model that vastly expands search capabilities in that growing medium.
Parsons notes that firms participating in the public beta have reported that AI search is driving improvements in overall knowledge management practices. Firms are refining and structuring their digital archives to make content more accessible, effectively creating a self-reinforcing cycle — as AI search improves, so does the quality of the knowledge it retrieves.
The Marketing Perspective: AI Search as a Strategic Advantage
For marketing teams, AI search can streamline content creation and proposal writing. Hillary Thompson, Director of Brand Communications and Senior Associate at MBH Architects, says before adopting AI-powered search, her firm’s marketing team spent hours manually pulling content for proposals and RFIs. Now AI quickly generates responses based on past projects, technical expertise and design approaches, improving speed and accuracy.
AI search also eliminates the need for marketers to know exactly where information is stored. In traditional systems, users struggle to locate content because they aren’t sure whether it’s in a project database, a proposal folder or a presentation archive. With AI, they can describe what they need in natural language and the system retrieves relevant information across all sources — including non-traditional sources like employee-authored posts and video transcripts.
As the firm grows more accustomed to AI search, it uses this tool to improve internal communications and storytelling, crafting stronger, more insightful narratives that differentiate the firm. “We’ve expanded AI search to help generate blog ideas, social media captions and content summaries, ensuring that our messaging aligns with our expertise,” says Thompson.
AI search also helps marketers and others “talk the talk” of the firm’s capabilities better and more independently. “We always tell clients how we’re experts, what makes us different and why they should use us over someone else,” says Thompson. “Now, when someone asks us a really technical question, we can find everything going on in a very evergreen way. We get the resources and techniques we’ve built around that topic just by typing in a prompt.”
One of MBH’s AI search requirements was that the function be limited to the firm’s knowledge only. “Our problem statement was, how do we access the collective brain of our company over time, and continually renew as things change rapidly, without explicit knowledge of where the information lives and who it lives with? We specifically didn’t want the product to query the internet; if it’s not our information, not something we’ve authored or that we’re using in the business, it’s inauthentic and purposeless to us in terms of marketing.”
The Technology Perspective: Unlocking Hidden Knowledge
Jim Martin, Vice President and CIO of 180-person, five-office architecture firm Shepley Bulfinch, says that their 10-year-old intranet originally stressed enterprise social content, but soon incorporated knowledge management practices as well. As the firm grew and expanded geographically, this latter piece became increasingly important.
“The quality of our intranet was always high, but initially it was more about discussions and connecting people,” says Martin, who led the rollout of “Finch,” the firm’s intranet. “Other firms invest more in curated content by experts, with a lot of pages written, and all the value that goes with it. We had some of that early on, but not as much as others did. We started building it once we realized its value.”
Martin says the Synthesis AI video transcription tool is a game changer for Shepley Bulfinch. “We have well over 400 videos on our intranet — we’ve said a lot of things in town hall meetings, quarterly firmwide addresses and learning and development courses that were never written down. But we record everything, so we have all these videos kicking around. Unless you were there, you wouldn’t know what was on that video, but now it’s all visible to AI search. It can summarize content that is spoken, not just written, then link to the exact spot in the video where that content is being discussed. So all that information is much more accessible than ever before.”
A major challenge of AI search is ensuring that information remains accurate and current. Older intranet entries may no longer reflect best practices, yet AI will still retrieve them as part of search results. How can firms curate their digital knowledge while benefiting from AI-powered retrieval?
Shepley Bulfinch is exploring ways to integrate AI search with content governance, ensuring that results highlight verified and current information while allowing users to cross-reference past insights. “It’s a good opportunity to reinforce the safe use of AI,” says Martin. “Trust but verify. AI accelerates knowledge access, but human expertise remains critical for correct interpretation.”
The Sustainability Perspective: A Vital Tool
AI search can be beneficial to sustainability initiatives as well. Corey Squire, Sustainability Director for BORA Architecture and Interiors in Portland, Oregon, says effective knowledge management is key to achieving high-performance outcomes. It can help firms develop design strategies, process improvements and communication best practices that contribute to resilience, health and equity goals.
Squire says, “There are certain questions that are asked again and again, like ‘what wall insulation should I use for this climate in this type of building’? And someone else has to stop and answer the question. Now, everything is documented and searchable. They can find what they need, and I can focus on other, more innovative things that move our practice forward.”
Squire has contributed data to BORA’s knowledge management collection by uploading to their intranet his entire book — People, Planet, Design: A Practical Guide to Realizing Architecture’s Potential — and making it searchable for employees.
The Broader Business Impact
Ultimately, AI-powered search enhances cross-functional collaboration by making firm-wide expertise more accessible (but with the capability to selectively limit who can see what). Parsons offers the recent case of a highly regarded, often-consulted senior engineer who meticulously documented his work. When he unexpectedly passed away, rather than losing all his knowledge and experience to history, the firm institutionalized his documented experiences so employees can still “consult” their former colleague for decades to come.
As AI-powered intranet search tools continue to evolve, their value extends beyond efficiency gains. They bridge knowledge gaps, improve collaboration and elevate the ability to make data-driven decisions. However, as Jim Martin points out, AI search is not a one-size-fits-all solution. AEC firms must adapt it to their unique workflows and knowledge management practices. Additionally, AI tools should be viewed as assistants rather than replacements for human expertise, ensuring that insights remain accurate, contextually relevant and strategic.
You can view the Synthesis AI Search landing page here: www.knowledge-architecture.com/synthesis-ai-search
If you are using AI Search in your firm, or you’d like to discuss any aspect of this subject, I’d love to hear from you. Write to me at rich@friedmanpartners.com or call 508-397-9213. Thank you for reading The Friedman File.
