LightTable, an artificial intelligence-driven peer review platform for real estate development, announced Tuesday that it has closed a $6 million seed round.

Generalist venture capital firm Primary Venture Partners led the round with early-stage investors Innovation Endeavors and MetaProp participating, as well as angel investors. Lighttable previously raised $2.2 million in a pre-seed round led by Primary and MetaProp.

SEE ALSO: Newmark’s Wayne Gearey: 5 Questions

Developed with input from architects and engineers, the Denver-based LightTable’s AI aims to expedite the construction industry’s peer review process by reducing a typically weeks-long process to well under an hour, said Paul Zeckser, co-founder and CEO at LightTable.

“We accept PDFs of design documents, drawings and specifications,” said Zeckser, who started the company in October 2024 with Dan Becker, now its chief technology officer, and Ben Waters. “The user will upload [those] to LightTable, and between 10 and 45 minutes later, we will have read through a couple of thousand pages of specifications. We will have done a full coordination of constructability, mechanical, electrical and plumbing engineering, accessibility, and fire and life safety.

“The reason that’s so revolutionary is because today there’s a three- to six-week process. You have to give those documents to somebody, and they’re going to take hundreds of hours looking at every single detail and checking the specs.”

After LightTable’s AI review, the company can create a prioritized list of issues for project stakeholders, highlighting what it believes will become costly [requests for information] and change-orders if not addressed.

“We are using all the available large language models and computer vision models, but then on top of that, we have built our own proprietary code base with guidance and expertise from architects, engineers and general contractors,” Zeckser said. “Our AI engineers are working very closely with our architects and structural engineers, creating an AI that is customized for this particular use case.”

LightTable claims that its AI identifies four times more issues than traditional peer reviews and reduces coordination mistakes on job sites by up to 70 percent.

“Our primary customer is a developer, but the developer shares it with everyone on the team, so architects and anyone involved in the design documents will use it,” Zeckser said “Once the developer buys it, everyone else on the project gets access to it for free for that project.”

The startup is working with some top multifamily developers, having signed a pilot agreement with Mill Creek Residential Trust, said Zeckser. 

“LightTable caught mistakes in seconds that outside experts spent weeks trying to find,” Shawn Poore, vice president of construction at Mill Creek, said in a statement. 

LightTable plans to use the seed funding to invest in more product engineering, along with experienced construction experts, architects, engineers and general contractors.

“​​Where we have gotten a lot of value is from the people that have been doing this for 30 or 40 years,” said Zeckser. “We’ve learned that peer review is not always black and white. It can be very nuanced. And having folks that have been doing it for a long time and have the battle scars, who then can help us understand why the nuance is important on a particular comment, that’s critical for us. They’ll tell us, ‘This comment is close, but here’s what the AI is missing.’ Once we understand that, we can take that back to our AI engineering team and they can improve it.”

Zeckser said that the marriage of field expertise with AI know-how is what separates his firm.

“We don’t think you can just build solutions with AI engineers who have never worn a hard hat or built or designed a building in Revit,” he said, referencing a popular architecture software. “You have to have those skills, and you have to really have mastered them to be able to tackle this space. We’re building a company that has deep expertise in the construction and architecture fields, but also in the artificial intelligence, large language model and computer vision fields.”

Philip Russo can be reached at prusso@commercialobserver.com.