Provision raises $7 million USD to build AI co-pilots for pre-construction estimates

Provision wants to help pre-construction professionals move faster and minimize errors with the help of artificial intelligence (AI).

Provision aims to build “the most accurate co-pilot for construction estimators.”

The Toronto-based construction technology startup has secured $7 million USD ($9.8 million CAD) in seed funding to develop its AI co-pilot, which helps construction estimators find information and identify risks and opportunities based on project documents.

The round, which closed in August, was led by Cercano Management with support from fellow new backer Jamie McDonald of Black & McDonald, as well as existing investors Y Combinator and One Way Ventures. It brings Provision’s total funding to $8.7 million USD.

Provision was founded in 2022 by CEO Luigi La Corte and CTO Brendan Ardagh, a pair of entrepreneurs with combined experience in construction, tech, and investing.

La Corte, the son of an immigrant general contractor, worked in civil engineering with Arup and infrastructure investing through Plenary Group. Ardagh—a quantity surveyor by training—formerly held the CTO role at human resources software startup Jem HR and co-founded student accommodation marketplace DigsConnect.

Provision claims that billions of dollars are lost every year due to errors that could have been caught at the pre-construction stage of projects, from design mistakes to scope misunderstanding and inaccurate estimates.

With Provision, La Corte and Ardagh aim to build “the most accurate co-pilot for construction estimators,” who must determine project costs by reviewing huge sets of complex documents. Provision aims to make those estimators’ lives easier by helping them answer targeted questions about documents quickly and accurately.

“We originally started to help reduce risk across projects, but we found that estimators had the biggest incentive to get this right during their process,” La Corte told BetaKit. “Originally we simply pointed out risks, and now we help pre-construction teams reduce the time it takes to prepare estimates.”

La Corte said that the company is developing its own models for specialized tasks that existing large language models are incapable of addressing effectively. The startup is betting that its domain-specific AI products—which include Project Review, Chat Agent, and Scope Agent—are better suited to tackling this challenge than more general-purpose chatbots.

In the construction industry, the cost of getting things wrong during the planning phase can be high, as severe or repeated miscalculations can lead to legal problems and business failures.

While Provision believes AI models can boost the productivity of construction estimators, there is also the risk of AI-generated responses presenting false or misleading information as fact, a problem that appears to be increasing with recent, more complex models.

“Given the high risk of miscalculation, we’re investing heavily in generating data sets to get up to 99 percent accuracy,” La Corte said. “This is very hard, since there are no datasets publicly available and building them is meticulous.”

La Corte said Provision’s long-term plan is to help companies “gut check any pre-construction work with multiple agents that work alongside estimators to ensure no mistakes are made.”

Provision’s traction has helped it attract investors. The company is already working with large construction firms in Europe and Canada like Acciona, Bird Construction, Colas, EllisDon, and PCL Construction, and it says it has reviewed $100 billion worth of projects after growing nearly 10x last year to more than $1 million in annual recurring revenue. The startup plans to add eight more employees to its 10-person team in the near future.

Feature image courtesy Freepik.