Compliance Software Dili Lands $21.7M as Federal Rules Collide With Infrastructure Spending

The Compliance Tax on America's Infrastructure Push

The wave of federal infrastructure spending - data centers, clean energy plants, advanced manufacturing - comes with a hidden surcharge: an ever-expanding rulebook. Davis-Bacon prevailing wage standards, apprenticeship ratios under the Inflation Reduction Act, OSHA site protocols, EPA permits - each layer adds complexity, and every misstep can trigger penalties that stretch into seven figures.

Dili, a compliance software startup founded by Anand Chaturvedi, closed a $15 million Series A round this week to turn that regulatory maze into a structured data problem. Khosla Ventures led the round, joined by Allianz, Rebel Fund, Brick and Mortar Ventures' Darren Bechtel, and Y Combinator's Garry Tan. Combined with a $6.7 million seed, the company has now raised $21.7 million since graduating from Y Combinator's summer 2023 cohort.

At DailyTechWire, we've tracked dozens of "AI for compliance" pitches over the past eighteen months, but Dili's angle is narrower and more defensible: it targets the intersection of federal funding, construction timelines, and labor law - a space where mistakes are expensive and manual audits are slow.

How the System Works

Dili's architecture splits the problem in two. Large language models handle the intake layer, ingesting unstructured documents from contractors, payroll systems, ERP platforms, and vendor submissions, then translating them into structured records. Once the data is clean, a deterministic rules engine - no probabilistic guesswork - compares it against the relevant statutes.

The result: tasks that previously consumed a full workday now finish in minutes. Chaturvedi offers Davis-Bacon wage verification as an example. The Department of Labor sets prevailing wages for federally funded projects, and a separate set of prevailing wage and apprenticeship rules applies to clean energy work under the IRA. Missing a threshold or misclassifying a worker can expose general contractors to fines that dwarf the cost of the software.

"Non-compliance can result in millions of dollars of fines for those projects," Chaturvedi noted. "So it's really powerful to be able to check all the information as it comes in, instead of just sampling data."

That claim hinges on reliability. Contemporary AI models are notorious for hallucination, and a compliance system that invents data is worse than useless. Chaturvedi's answer is architectural: LLMs touch only the parsing layer, never the decision logic. The rules themselves - complex but static - are encoded deterministically, so the system either flags a violation or it doesn't. There's no room for a model to "creatively interpret" a regulation.

Seven Hundred Projects and Two Business Models

Dili is live on approximately 700 construction sites, spanning manufacturing facilities, data centers, and energy infrastructure. The split is roughly even between two deployment models: half the customers use Dili as in-house software, giving their own compliance teams a faster workflow; the other half outsource the entire function to Dili on a contractor basis, effectively buying compliance as a service.

Chaturvedi sees the balance shifting over time. As AI tools mature and construction firms grow more comfortable with software-driven workflows, he expects more customers to bring compliance in-house rather than pay for managed services. "Software and AI are going to start eating a lot of those professional services workflows," he said. "The interesting thing will be how the market itself evolves and where the customer needs go as AI develops."

That evolution matters for Dili's unit economics. Software margins are higher than services margins, and a customer base that skews toward self-service will scale more cleanly. But the dual model also hedges risk: firms that lack internal compliance expertise can still buy in, and Dili collects revenue either way.

Why Khosla and Bechtel Are Betting on Compliance

The investor roster tells the story. Khosla Ventures has a long track record in infrastructure-adjacent software, and Allianz brings insurance-industry perspective on risk and liability. Darren Bechtel, through Brick and Mortar Ventures, represents the construction establishment - a signal that incumbents see software as a necessary upgrade, not a threat.

The timing is also deliberate. Federal infrastructure spending is accelerating, and the regulatory surface area is growing in parallel. The IRA alone introduced new labor standards tied to tax credits, and the CHIPS Act layered on additional requirements for semiconductor fabs. Every new program adds pages to the rulebook, and every page is a potential landmine for contractors who rely on spreadsheets and spot checks.

Dili's pitch is that compliance should be continuous, not episodic. Instead of sampling payroll records at the end of a quarter, the system ingests data in real time and flags discrepancies before they compound. That shift - from audit to monitoring - is the real product innovation, and it maps cleanly onto the capabilities of contemporary AI.

The Limits of Deterministic Compliance

Chaturvedi's architecture is conservative by design, and that conservatism is a feature. But it also exposes a limitation: the system is only as good as the rules it encodes. When regulations are ambiguous or subject to interpretation - and labor law is famously gray in places - a deterministic engine can't resolve the uncertainty. It can flag the edge case, but a human still has to make the call.

That's fine for now. Dili is selling speed and accuracy on the 95 percent of cases that are straightforward, and leaving the hard calls to compliance officers. But as the product matures, customers may push for more guidance on those edge cases, which could pull Dili back toward probabilistic models or, more likely, toward building a knowledge base of past rulings and precedents.

The other constraint is adoption friction. Construction is not a software-first industry, and getting contractors to feed clean data into a new system requires change management and, often, integration work with legacy payroll and ERP platforms. Dili's dual business model - offering both software and managed services - is a hedge against that friction, but it also means the company has to maintain two operational playbooks.

What Comes Next

With $21.7 million in the bank, Dili's immediate priorities are likely headcount and customer acquisition. Seven hundred projects is a strong start, but the addressable market - every federally funded construction site in the United States - is orders of magnitude larger. Scaling will require both sales capacity and product refinement, especially around integration and user experience.

Longer term, the question is whether Dili can expand beyond federal compliance into other regulatory domains. State and local rules vary widely, and private projects have their own insurance and contractual requirements. If the core engine - LLM parsing plus deterministic logic - proves robust, the same architecture could apply to permitting, environmental review, or safety reporting.

For now, though, the company is riding a tailwind. Infrastructure spending is up, regulatory complexity is up, and the cost of non-compliance is up. Dili's bet is that construction firms will pay to offload that risk, and Khosla's bet is that the market is big enough to build a category leader. If both are right, we're watching the early innings of a new software vertical - one that emerged not from Silicon Valley's product imagination, but from the grinding reality of federal rulemaking.