Parspec, a startup using artificial intelligence to improve efficiency in the construction industry supply chain, has raised $20 million in a Series A funding, the company told Crunchbase News exclusively.

The company aims to help sales agents and wholesale distributors “efficiently bid and supply” construction products. The company claims its key differentiator is its ability to “instantly identify products available in the market that satisfy complex specifications provided by the customer” using artificial intelligence.

“This enables customers to cut time and cost to bid in half, while simultaneously improving bid quality and compliance, enabling them to bid and win more projects,” CEO and co-founder Forest Flager said in an interview.

The company began with a focus on lighting and electrical products and has since expanded to support mechanical, electrical and plumbing (MEP) products.

Threshold Ventures (formerly DFJ) led Parspec’s Series A financing, with participation from existing backers including Innovation Endeavors, Building Ventures, Heartland Ventures and Hometeam Ventures. The funding brings San Mateo, California-based Parspec’s total raised to date to $31.5 million.

While the company declined to reveal its valuation, Flager told Crunchbase News that it was “up 3x” from the time of its seed raise in March 2022.

The raise comes at a time when venture funding to real estate-related startups in the U.S.real estate funding in the United States has been on the decline. The sector saw a total of $13.7 billion raised across 1,088 deals in 2021, per Crunchbase data. Hurt by a surge in mortgage interest rates, property prices and construction costs, real estate startup funding plunged to $4.9 billion in 2023 before dropping even further to $3.6 billion for 2024. So far in 2025, $2.1 billion has flowed into real estate startups across 201 deals in the U.S.

Optimization through information

Flager met co-founder Pratyush Havelia while he was working on his post-doctorate degree at Stanford University. Havelia was a student in his research lab. In conducting research into how designers could leverage computing to better explore options and optimize buildings for things like energy efficiency, the pair realized that the “optimizations are only as good as the information you feed in,” Flager said.

“Information about building products, such as the performance of different windows or lighting materials, was not really available at scale,” he added. “It was just really hard to get that information.”

Flager went on to lead the software team at Katerra, a SoftBank-backed construction tech company that ultimately went under in 2021. While there, he gained a better understanding of the kind of mechanics of how material gets purchased, and also the sort of information the distributor has.

“I wanted to build a solution that would empower the distributor, and ultimately provide a means to collect product information in a way that I thought can be useful to many different players in the value chain,” he recalls.

Parspec was incorporated in 2020, but Flager and Havelia began focusing on it full time in 2021.

Flager claims that today there are no other products on the market that are able to instantly identify spec-compliant products or automatically locate product documentation such as install guides or warranties. That differentiator, in conjunction with the launch of its quoting product in January 2024, gives the company a competitive edge, he believes.

As a result, Parspec has experienced 4x revenue growth over the past 12 months.

Flager believes that Parspec — which combines the words “parse” and “specifications” — has the ability to help drive down construction costs across the supply chain. Construction material prices are approximately 40% higher than they were in 2020, so to Flager a more agile and efficient construction supply chain is needed more than ever.

AI component

Describing Parspec as a “workflow tool,” Flager said each component in the workflow leverages AI capabilities. The company uses machine learning and LLM models to pull out products from unstructured documents such as drawings and specifications. It then identifies the technical requirements from the natural language spec.

“Then, when we’re talking about matching products to spec, we have the whole AI pipeline to create comprehensive pipeline where we’re crawling about 4,000 different manufacturer websites, indexing all the PDF product documentation they have there, and then using AI ML models to pull out the attributes and create that catalog that matches those products to the requirements,” he added.

Parspec currently has 288 customers, primarily wholesale distributors and sales agencies in the MEP product verticals. The company said it’s working with four of the five largest electrical distributors and three of the five largest lighting agencies in the U.S.

Its revenue model is a usage-based subscription. The usage component of the subscription is based on the number of documents the customer creates using the Parspec platform. Qualified documents include quotes, submittals, and operation and maintenance packages.

Threshold Ventures partner Mo Islam told Crunchbase News via email that his firm was impressed with Parspec’s “truly AI-native product to automate the procurement process,” which he believes allows the company’s customers “to win more business while improving profit margins in a trillion-dollar-industry.”

“Parspec’s proprietary materials dataset and fine-tuned generative AI models give the company a strong data moat and AI competitive advantage,” Islam added. “The company’s platform can automate the entire procurement workflow for the construction industry, from product selection and pricing to quoting and submittals.”

Related Crunchbase query:

Related reading:

Illustration: Dom Guzman