Dili raises $21.7 million to use AI for the paperwork nightmare of America's construction boom
A startup called Dili has secured fresh funding to help infrastructure projects, think data centres and factories, stay on the right side of the overlapping federal rules that govern wages, safety and the environment.
Key points
- Dili raised a $15 million Series A round, led by Khosla Ventures, bringing total funding to $21.7 million.
- The startup's software currently runs compliance checks across roughly 700 construction and infrastructure projects in the United States.
- Dili targets federal rules such as Davis-Bacon prevailing wage requirements and clean-energy apprenticeship rules under the Inflation Reduction Act.
- The company uses AI only to read and sort raw documents, then hands the results to a traditional rule-based system to reduce errors.
- About half of Dili's customers run the software themselves; the other half pay Dili to manage compliance end-to-end.
Building a data centre or a clean-energy plant in the United States means wading through a thicket of overlapping federal rules. Get something wrong and the fines can run into the millions. A startup called Dili thinks AI can do the wading.
Dili announced this week that it raised $15 million in a Series A round led by venture firm Khosla Ventures. Allianz, Rebel Fund, investor Darren Bechtel of Brick and Mortar Ventures, and Y Combinator president Garry Tan also joined the round. Dili previously went through Y Combinator's Summer 2023 batch and had already raised a $6.7 million seed round, bringing its total capital to $21.7 million.
What rules are we actually talking about?
Three main bodies of law create the compliance headache Dili is trying to solve. The Davis-Bacon Act lets the Department of Labor set minimum "prevailing wages" for workers on federally funded construction projects. Separately, the Inflation Reduction Act, the 2022 climate and energy law, added its own prevailing wage and apprenticeship requirements for clean-energy projects. On top of those, OSHA (the Occupational Safety and Health Administration, the federal workplace-safety regulator) and the EPA (the Environmental Protection Agency) each have their own rules depending on what is being built.
For a single large project, those rules can create thousands of documents and payroll records that all need checking.
"Non-compliance can result in millions of dollars of fines," Dili co-founder and CEO Anand Chaturvedi told TechCrunch AI. "So it's really powerful to be able to check all the information as it comes in, instead of just sampling data."
How does Dili actually use AI?
Dili is careful about where AI does the work. The company uses a large language model, the technology behind chatbots like ChatGPT, only at the document-reading stage. The model converts messy, unstructured files, scanned payroll sheets, vendor invoices, internal reports, into clean, structured data that a computer can work with.
After that, a conventional, rule-based system takes over. It applies the compliance rules, which are complex but fixed, without any AI guesswork. Chaturvedi says a task that once took a full day can now finish in minutes.
Dili says its software is already live across roughly 700 projects, from manufacturing plants to data centres.
| Milestone | Detail |
|---|---|
| Seed round | $6.7 million |
| Series A round | $15 million (2025) |
| Total raised | $21.7 million |
| Active projects | ~700 |
| Y Combinator batch | Summer 2023 |
What does this mean for the people doing compliance work today?
Right now Dili splits its business roughly in half. About half its customers use Dili as an internal software tool; the other half outsource their entire compliance function to Dili directly, much as a company might hire an outside accounting firm.
Chaturvedi expects that balance to shift. "Software and AI are going to start eating a lot of those professional services workflows," he said, predicting more companies will bring compliance work in-house over time.
For workers in compliance and construction administration, that shift is worth watching. It does not mean immediate job cuts, but it does suggest the role will change, with less time spent reading documents by hand and more time reviewing what the software flags.
Common questions
Does this affect me if I work on a construction site?
Not directly. Dili's software is used by project owners and contractors to check that wage and safety rules are being followed, which should actually help workers get paid correctly under federal law.
Is AI making legal compliance decisions here?
No. Dili uses AI only to read and organise documents. A traditional, rule-based system makes the actual compliance calls, which limits the risk of AI errors affecting legal outcomes.
