Artificial intelligence is already changing construction, but one of the most adopted uses of the technology comes from Document Crunch, an Atlanta-area startup that’s now automating one of the more time-consuming upfront tasks in the construction process: lawyers and project managers sifting through hundreds of pages of contract documents to find language that increases risk to a contractor or client. Document Crunch co-founder and CEO Josh Levy did this work as an attorney at contractors JE Dunn and Wood before connecting with tech entrepreneur and CTO Trent Miskelly to apply a large language model to this repetitive task.
“It’s becoming more and more apparent to me that the brand, the way you go to market, and the empathy that’s in our DNA for construction, serve to be a differentiator,” says Levy. “This incredible amount of investment in technology, the core technology that is infused with the subject, is significant.”
Risk reduction of project documents
Photo courtesy of Document Crunch
Document Crunch’s machine learning platform uses construction contract-specific models and generative AI frontier models from OpenAI and Anthropic, as well as a retrieval architecture designed to provide answers to complex contractor questions. It also has a chat function to allow users, often internal advisors, to ask questions about specific projects. Natural language AIs are becoming increasingly popular for allowing users to essentially talk to the AI. The company’s platform has been adopted by ENR’s top 500 contractors, realizing that while contracts cannot be standardized, contractors can analyze them.
“We’re seeing things before most others [do] and the crazy thing is, every time I get debated with an ENR company, I’m sitting in a boardroom with the CIO, the CEO, and people who basically make decisions for the company, and they’re diving deep. about how they will change the [firm’s] trajectory, taking advantage of these new products and new technologies,” says Miskelly.
Applying knowledge from a database full of construction experience from Levy and other construction legal experts not only saves time for in-house counsel and other attorneys when assessing contractor risk, but which also frees them up for more important work.
“This [document review] it’s a repetitive process, in some ways, asking yes or no, but there’s an art to it,” says Jeff Brannen, Texas and Arizona legal director for contractor Balfour Beatty. “Understanding and knowing, having the vision and to say that computer software is in a place where it can now do that, it can help us.”
Document Crunch raised $21.5 million in a Series B funding round last October led by Titanium Ventures, with participation from Nemetschek AG. Levy said the company plans to use the funding to expand its various document review AI agents into a full platform game that could even go beyond construction documents. The company last July launched its Construction Contract Checklist and Project Playbook, tools that are designed specifically for on-site personnel such as site crew managers, site managers, projects and project engineers.
Yash Patel, general partner at Titanium Ventures who joined Document Crunch’s board with the new funding round, said developing products to help clients address the risk profile at any stage of a project is a focus for 2025.
