Buildcheck raises $12m Series A for AI-powered drawing review

The Stanford-founded construction AI startup has raised $12m, bringing total funding to $18m, to expand its vision-based drawing review platform.

Buildcheck raises $12m Series A for AI-powered drawing review

Buildcheck AI has closed a $12 million Series A led by Telescope Partners, with participation from WND Ventures, the corporate venture arm of DPR Construction, alongside returning investors Uncork Capital, Salt VC and Xfund. The round arrives nine months after a $5.9 million seed, taking total funding to $18 million. The San Francisco company says proceeds will fund expansion across general contractors, developers and design firms, and the launch of four new product modules.

The platform applies computer vision models to construction drawing sets, running automated checks designed to catch errors, omissions and coordination conflicts before they reach the field. Co-founder and chief executive Alexander Michalatos said Buildcheck has analysed more than 150,000 sheets to date and grown its customer base more than fourfold over the past year to over 110 paying customers. Named clients include DPR Construction, EllisDon, Power Construction Company, EMJ Construction and IMC Construction.

What the funding buys

Buildcheck is using the round to ship four product extensions that it says move the platform from a point-checking tool to a broader design review environment. Custom Checks allows firms to upload their own QA/QC manuals, which the platform converts into automated review agents enforcing internal standards across every drawing revision. Diffs provides a visual and plain-language comparison between drawing versions, with logic calibrated to surface substantive changes rather than geometric shifts caused by formatting. Two further modules, covering code compliance and value engineering, are entering beta for customers in the US and Canada.

Chris Gaertner, Principal at Telescope Partners, said the quality of the customer base and reported return on investment were what drove the firm's conviction: "Drawing review has been an accepted pain for decades, costing the industry billions. Buildcheck has changed that."

Zach Murphy, Design-to-Build Leader at DPR Construction, said a pilot with Buildcheck was helping its reviewers prioritise critical issues by automating the initial pass over design documents.

Market context

Construction technology remains one of the larger enterprise software opportunities yet to see full platform consolidation. The source release cites an industry-wide figure of $200 billion in annual costs attributable to design errors in a $13 trillion global construction market, though those numbers are not independently attributed. Several well-funded startups are pursuing AI-assisted preconstruction workflows, including document management, clash detection and specification compliance, alongside established players such as Autodesk and Procore that have embedded AI features into broader project management suites.

Buildcheck's positioning targets the preconstruction phase specifically, where errors are least expensive to correct, and the company reports customer ROI of 10 to 35 times. Those figures are self-reported and unaudited, and the release does not disclose average contract value, annual recurring revenue or customer retention rates. Investors and buyers will watch for independent benchmarks and reference data as the company scales beyond its current base.

The construction sector's AI adoption is also being shaped by a broader push toward interoperability standards, with bodies such as buildingSMART International advancing open data formats for building information modelling. Compliance with those standards could influence how drawing-review AI tools integrate with project delivery platforms at scale, and Buildcheck has not stated its current level of BIM interoperability.

Buildcheck was founded at Stanford University by Michalatos, Andrei Molchynsky and Alex Gureiev, who serves as CTO and led development of the company's proprietary computer vision models.