This September, AVAIL hosted 50 of the AEC industry’s top technology leaders at Confluence Lexington 2024. We spent three uninterrupted days exploring the progress AEC industry firms and technology companies have made in implementing artificial intelligence (AI) and machine learning (ML) technologies thus far, as well as the impact they could have on the industry in the next 12 to 18 months.
Our practitioner speakers hailed from leading firms Olson Kundig, Perkins&Will, Thornton Tomasetti, Wade Trim, and SHoP Architects and shared how these innovative AEC firms are thinking about AI and ML and putting the technologies into practice. We have distilled their session insights below, highlighting commonalities that can be easily implemented by likewise forward-thinking firms to create effective AI and ML strategies.
Jump to
| At Confluence Lexington 2024, 50 AEC innovators gathered to discuss Progress in AI and ML, A Year Later. | Confluence attendee Natalia Locke from building, engineering, and design firm Introba networks with other AEC technology leaders over lunch at Confluence Lexington. |
Given the theme, Progress in AI and ML, A Year Later, our Confluence Lexington 2024 practitioner speakers gave our group of 50 AEC innovators a lot to think about in relation to artificial intelligence (AI) and machine learning (ML). Here’s a recap of their sessions.
Kenton Grant, Director of Technology at Olson Kundig
At Confluence Lexington 2024, Olson Kundig Director of Technology Kenton Grant described the firm’s exploration and integration of AI technology within their design process. The firm has taken a phased approach, moving from informal experimentation by individual staff members, to more structured use of AI on select client projects, and now towards a committee-driven strategy for ethically and effectively deploying AI firm-wide. Kenton emphasized the importance of play in figuring out the beginning applications of AI. Now, the firm aims to empower this play while maintaining strategic guidelines around AI and ML initiatives.
Charles Portelli, Associate | Digital Innovation Strategist at Perkins&Will
Perkins&Will Associate | Digital Innovation Strategist Charles Portelli discussed the potential of AI to assist designers. The goal at Perkins&Will is to build "intelligence" by combining data, information, and the experience of designers. The firm has developed tools to clean and organize data from architectural models and projects, in order to create a knowledge base that can support designers with decision-making and design tasks. The ultimate aim is to create an "informational assistant" that can augment designers' experience and decision-making, without replacing the core human expertise and judgment required in the design process.
Learn more about Perkins&Will's efforts to apply AI to their architectural practice in The Confluence Podcast's interview of Charles Portelli.
Robert K. Otani, PE LEED AP, CTO and Managing Principal at Thornton Tomasetti
Chief Technology Officer and Managing Principal at Thornton Tomasetti, Robert Otani discussed how the firm has embraced and integrated AI and ML into their workflows and processes. Thornton Tomasetti’s applications development, advanced computational modeling, and R&D group CORE studio has built an end-to-end infrastructure for developing and deploying these AI applications, which has led to the development of numerous AI-powered applications to automate and accelerate common engineering tasks such as structural analysis, connection design, and material quantification.
Rob Sinclair, Corporate Technology Lead at Wade Trim
Corporate Technology Lead for Wade Trim, Rob Sinclair, discussed the firm’s formation of an AI Task Force to explore and implement AI solutions in their business. The Task Force’s current key AI initiatives are a hydraulic modeling toolkit, an AI-powered proposal system, and a custom chatbot. Wade Trim is partnering with tech companies like YegaTech—who ran the Confluence AI Workshop in 2023—to help establish the firm’s implementation strategy regarding AI, and they are also engaging with universities to bring in AI talent.
John Cerone, Principal at SHoP Architects
SHoP Architects Principal John Cerone discussed how the firm has been exploring an industrialized, manufacturing-focused approach to design and construction. This includes using digital modeling, automation, and offsite prefabrication to improve efficiency and reduce errors. John also talked about SHoP's custom modular construction company, Assembly, and their efforts to further industrialize the building process involving digital template and workflow development to enable distributed supply chains and rapid assembly.
What is the history of Confluence?
Find out in our Q&A with host Randall Stevens.
Synthesizing the sessions from Confluence Lexington 2024, here are two easy steps AEC industry firms can take to create effective strategies around artificial intelligence (AI) and machine learning (ML).
Firms like Olson Kundig, Wade Trim, and Thornton Tomasetti have appointed leaders to think strategically about what AI projects the firms will prioritize, as well as to develop infrastructure and policies around the firm’s use of these technologies.
Olson Kundig’s AI Stewardship Group, for example, meets monthly and helps guide the strategic use of AI, addressing concerns around confidentiality and data governance and developing policies. What started as a bottom-up, cultural interest in playing with AI morphed into an appointed AI Stewardship Group consisting of representation from legal, marketing, design technology, design, and ownership. Now, the group is streamlining into a smaller committee that will be fed by AI focus groups, where staff can drive the AI and ML projects they think should be worked on.
“I see this as a top down, bottom up approach,” Olson Kundig’s Kenton Grant said in his Confluence session. It is top-down because there is a directive from his executive leadership to work with AI and ML, and it is bottom-up in the sense that this all started from explorative play from the design staff. “We want to make sure that designers still have that agency and the autonomy to be able to do those things in a more empowered, structured way,” he added.
Wade Trim’s AI Task Force also meets monthly to identify AI applications, assess feasibility and risks, and make recommendations. The firm has partnered with Confluence 2023 speakers YegaTech's Mehdi Nourbakhsh, Ph.D. and Sam Zolfagharian, Ph.D. to inform the AI Task Force and help strategize AI implementation across the firm.
AEC leaders should consider setting up similar cross-functional committees and teams to provide oversight and direction on AI initiatives.
Leading development at Thornton Tomasetti, the firm’s CORE studio—founded by Confluence speaker Robert Otani—is a fully integrated applications development, advanced computational modeling, and R&D group within the firm.
The infrastructure built out in CORE studio has been key in supporting custom AI and ML development that align with Thornton Tomasetti’s business goals. “We've built out the end-to-end infrastructure of how to go from ideation, to synthetic data generation, to training, and then to deployment,” Robert said in his Confluence session.
Perkins&Will has a similar integrated team within the firm. This structure helps Thornton Tomasetti and Perkins&Will scale the use of AI across the organizations.
In an effort to prepare for inevitable AI-driven changes, it is important to keep all teams up to speed. Firms like Thornton Tomasetti, Olson Kundig, Wade Trim, and Perkins&Will have prioritized firm-wide education on AI and ML initiatives as a way to proactively prepare their firms for the impact of AI.
Thornton Tomasetti’s Robert Otani emphasized the importance of prioritizing transparency and upskilling the existing workforce on AI concepts and applications. It is imperative to get buy-in and adoption from engineers, he noted. Engineers want to understand and validate the AI-generated outputs. He recommended firms invest in training and education to demystify AI and show how it can enhance rather than replace human expertise, as well as provide documentation, access to datasets, and ways for engineers to inspect and override the models if needed.
Charles Portelli from Perkins&Will mentioned the same. As AI becomes more prevalent, leaders should proactively address concerns and resistance from designers by providing training, education, and opportunities to experiment with AI in a low-stakes environment.
Likewise, Olson Kundig has created tutorial videos, lunch-and-learns, and other knowledge sharing initiatives to educate staff on using AI tools, and Wade Trim provides quarterly AI-focused town halls and training for employees.
AEC leaders should invest in similar training programs to get their teams up to speed and comfortable with AI.
| Confluence attendee Darya Montazerinamin and session speaker Rob Sinclair from Wade Trim listen to day one sessions at Lexington’s Fasig-Tipton. | For day two sessions, Confluence Lexington 2024 attendees traveled to New Haven’s Log Still Distillery. |
The possibilities of artificial intelligence (AI) and machine learning (ML) technologies are just starting to come into focus. They can be applied to a range of project types, and it is important that firms align on what projects to prioritize.
Before deciding what types of AI and ML projects to take on, it is important to first align on what you don’t want to lose in activating these technologies: human expertise.
A key theme of the Kenton Grant’s Confluence session was Olson Kundig’s need to preserve human involvement and responsibility, rather than fully automating processes with AI. He asked, “Where does AI belong? And where does AI not belong?” and recommended AEC technology leaders maintain human oversight and accountability in whatever projects they decide to pursue.
Perkins&Will’s Charles Portelli also spoke to this. He stressed the importance of preserving the human experience, design thinking, and decision-making in the design process. AI should be used as a tool to assist and empower designers, not to replace them. Thus, he recommended firms explore AI projects for specific use cases that can augment human capabilities, such as image generation, data normalization, and knowledge retrieval.
Below are recommendations from our Confluence Lexington 2024 speakers for the types of AI and ML projects AEC industry firms should consider prioritizing.
Before heavily investing in AI and ML applications, foundational work should be a priority, Charles stressed. He highlighted the importance at Perkins&Will of cleaning and normalizing data, understanding relationships between design elements, and building a knowledge graph to power reasoning and inference. It is imperative to have confidence and control over AI systems, given concerns around hallucinations and biases in AI models.
John Cerone from SHoP Architects expanded this thought. In his session, he recommended firms focus on getting data and digital workflows in order first before trying to apply AI and ML technologies. John emphasized the importance of having structured, accessible data from their 3D models and other digital assets as a prerequisite for meaningful AI applications.
It is easy to fall into the AI hype. Charles recommended firms approach AI cautiously. Leaders should take a measured, pragmatic approach, understanding the current limitations of AI and not overpromising what it can deliver.
An easy way to apply AI pragmatically is to think through ways it can help alleviate your team of repetitive work. Across the board, Confluence Lexington speakers emphasized the opportunity in architecture and engineering firms to automate repetitive, time-consuming tasks with AI and ML.
SHoP Architects’ John suggested exploring AI for automating tasks that are currently very manual like quantity takeoffs, scheduling, and other data-driven analysis.
Thornton Thomasetti’s Robert Otani showed how AI can be used to quickly size structural elements, design connections, and perform other routine engineering analyses. Targeting these low-hanging fruit can provide quick wins and demonstrate what AI can do for firm leadership.
Robert additionally recommended developing a strategy for integrating AI models into existing workflows. Rather than replacing engineers, the goal should be to augment their capabilities by embedding AI-powered tools and features into the software and applications they already use.
Rob Sinclair’s session touched on the exploration of AI-powered tools and solutions, as well. He highlighted several AI-powered tools and solutions Wade Trim is using, like proposal generation, hydraulic modeling, and utility mapping.
AEC technology leaders should identify opportunities to apply AI to improve efficiency and productivity across their operations.
Sometimes it’s best to start small. Rather than trying to broadly apply AI across the organization, Kenton's firm Olson Kundig identified targeted use cases like conceptual design, marketing materials, and client presentations. AEC technology leaders should take a similar approach, focusing first on discrete applications that can demonstrate the value of AI and ML.
Charles’s session touched on this as well. The key is to take a thoughtful, incremental approach to AI integration, focusing first on building the necessary foundation, and then carefully selecting use cases that can truly enhance the design process and the work of architects and engineers.
John suggested the same. Focusing on tangible use cases that can deliver value and building the necessary data foundations to support more advanced AI applications over time will serve firms well. He suggested keeping an open, experimental mindset will be important as the technology continues to evolve rapidly.
Photo credit: Steve Bennett, Taylor Design
AVAIL hosts Confluence events and The Confluence Podcast to understand what AEC industry firms are needing, wanting, and thinking around technology. These conversations feed our product development and strategy, and we are happy to continue them in-person and on the podcast in 2025.
We are planning a one-day event in San Francisco in the spring of 2025 and our signature three-day, invite-only Confluence Lexington event in the fall of 2025.
Register interest in joining us.
Explore The Confluence Podcast episodes.
Confluence Lexington 2024 was made possible thanks to the support of our technology partners AVAIL, Dell Technologies, NVIDIA, Intel, ArchVision, Egnyte, and Chaos.
Hosted by software company AVAIL, Confluence is a series of architectural, engineering, and construction (AEC) industry professional development events that bring design technology and product management leaders together in collaborative discourse. Its signature conference, Confluence Lexington, is a three-day, invite-only event themed around a pressing industry technology issue and held in the fall every year. In the last few years, regional, one-day events have also been added, as well as a podcast hosted by AVAIL CEO and Founder Randall Stevens and TRXL Media Director Evan Troxel.
Founded in 2016 as a solution for architectural and engineering firms to find the information they need faster, AVAIL creates software for the global architectural, engineering, and construction (AEC) industry. AVAIL’s content management system (CMS) platform and Revit application tools help designers and engineers take control of their intricate network of files in one visual, streamlined solution. AVAIL’s customers include industry leaders like Gensler, Perkins&Will, IMEG, LEO A DALY, and Populous.