Market Study · NYC Building Compliance
Roll up NYC’s fragmented compliance-services market — and run it on AI.
NYC building-code compliance is a large, legally-mandated, recurring-revenue services market split across 3,742 mostly owner-operated firms with no consolidator. The thesis: acquire a set of these shops at small-business multiples, keep their licensed relationships, and modernize the paper-heavy workflow with AI agents to lift margin and throughput per licensed head.
01 · Executive summary
The thesis in one paragraph
Every building in New York must stay compliant with a dense stack of building, safety, and housing codes; when it doesn’t, three city agencies issue violations that block permits, sales, and refinancing. A specialized services industry — permit expediters and code consultants — exists to clear them. Demand is large, growing, legally coerced, and recurring. Supply is extraordinarily fragmented: 3,742 firms, the top 5 holding only 11%, a median firm with just 7 lifetime filings. That combination — durable recurring demand + a long tail of aging, sub-scale, owner-dependent operators — is the classic setup for a buy-and-build roll-up. The differentiated angle is AI: the labor-intensive middle of the work is structured paper-shuffling that agents can automate, so a consolidated, software-enabled platform can run at materially higher margin than the shops it buys.
Fragmentation makes targets cheap and plentiful; you acquire the one scarce asset — agency relationships, licensed staff, and a book of recurring clients — that a new entrant cannot build fast.
The clerical core (read notice → map to form → file → chase hearing) is exactly what AI agents do well, so a modern operator lifts margin and throughput per licensed head.
Multiple arbitrage: buy sub-scale shops at ~3–5× SDE, integrate and modernize, and re-rate the assembled platform at a higher multiple.
02 · The opportunity
A services business, not a construction business
Expediters and code consultants are hired to get work approved by the Department of Buildings and to clear violations at DOB, ECB/OATH, and HPD. They interpret notices, file the correct paperwork, navigate agency systems, and represent owners at hearings; the physical repair is subcontracted. The scarce assets are relationships, process knowledge, and licensed staff — high-margin, low-capital, and sticky.
03 · Market size
Measured, estimated, and validated
The city has imposed $3.14B in penalties on owners all-time, a $144M/yr run-rate — summed directly from the data. An independent OATH ledger corroborates the imposed figure to ~1–2%/yr and shows NYC actually collects ~$83M/yr on a settled basis. Expediter revenue is a fee-sensitive estimate of $185M–$615M/yr on the job-filing slice alone. All of these are a floor on what owners spend — they exclude cure, legal, and expediter fees.
04 · Market structure
Nobody owns it
3,742 firms split 711,643 filings; the top 5 hold 11%, the top 25 only 31%. The mass of firms sits in the small, owner-operated band — the acquisition-target universe — with no natural consolidator. Even the largest player, a PE-backed roll-up, has barely moved the curve.
05 · Demand durability
The revenue comes back
A roll-up only works if revenue recurs. It does: 41% of all filings are on a property the firm already served, and 55% are for a client (managing agent) it already had. The established leaders run 65–80% repeat-client work — an annuity, not project churn. Aging housing stock and legally-mandated, deadline-driven violations guarantee the demand keeps arriving.
And the demand is coerced and escalating: an unresolved violation doesn’t sit still. Miss an OATH/ECB hearing and you owe the full fine plus interest plus a default penalty; many hazardous conditions accrue a daily civil penalty until cured; and unpaid judgments become tax liens on the property that can freeze a sale or refinance. The cost of not hiring someone rises every day — which is exactly what makes this non-discretionary spend.
06 · Competitive landscape
Who the players are
The visible leaders range from a private-equity-backed consolidator (Milrose) and a 50-plus-person independent (Outsource) to a 1969-era code house (William Vitacco) and a six-person shop pushing 17,000 filings. Their profiles both illustrate the archetypes to buy and set the benchmark a modernized platform would beat.
07 · The play
Roll-up + AI enablement
Acquire 5–10 owner-operated shops at small-business multiples, preserving the licensed relationships and recurring client books that can’t be rebuilt. Centralize the back office and OATH hearing representation, cross-sell violation-cure ↔ permits, and use the public violation feeds to outbound-sell owners the day a violation lands. Then apply the wedge:
Buy the moat
Agency relationships, licensed staff, and a returning client book — cheap and plentiful because the market is fragmented and aging.
Automate the middle
AI agents parse notices, pre-fill and queue filings, flag missing evidence, and monitor feeds — turning clerical hours into review-and-approve.
Re-rate the platform
Higher margin and throughput per licensed head, assembled into a scaled platform worth a higher multiple than the sum of its parts.
08 · Precedent transaction
This exact playbook already worked once
The thesis isn’t hypothetical — the market’s largest firm is a completed PE roll-up. In 2019 lower-middle-market PE firm Southfield Capital partnered with Milrose Consultants, then a NYC-focused code and permit-expediting shop. Over the hold they executed 14 add-on acquisitions, expanded from New York to offices in all 50 states and 435+ professionals, and grew earnings more than 7×. In June 2026 Southfield sold the platform to a larger PE firm, Littlejohn & Co. — a clean fragmentation → consolidation → PE-exit arc, in this precise vertical.
The model also travels beyond NYC: permit expediting and code consulting is a national service category in every dense, heavily-regulated metro. Milrose grew from a NYC shop to offices in all 50 states, and Chicago-founded Burnham Nationwide runs the same “expediting + in-house code consulting” model across Chicago, LA, San Jose, New York, and Denver. NYC is the densest, most-fragmented beachhead; the same playbook ports to Chicago, LA, SF, Boston, DC, and Miami once the platform is built.
What we would do differently — and what makes the return profile better today — is layer AI enablement onto the roll-up. Milrose scaled primarily by adding people and offices; a second-generation consolidator can automate the clerical core and compound margin per licensed head on top of the same acquisition engine.
Sources: Southfield Capital and Littlejohn & Co. press releases (2019 partnership; June 2026 sale); Burnham Nationwide (company offices). Deal terms were not disclosed.
09 · Risks & mitigants
What would break the thesis
Keyman / succession
Value can walk out with a retiring founder. Mitigant: earn-outs, retention, and transferring relationships onto the platform during a transition period.
Commoditization
The city wants self-service filing, so simple filings may commoditize. Mitigant: concentrate on complex violations and hearings — and be the automation, not its victim.
Enforcement history
The category has a history of DOB bribery scandals. Mitigant: diligence each target’s enforcement record; compliance-first operating model.
Collection risk is real too: imposed penalties are ~2× the cash the city collects, so revenue-cost figures should key off collected, not imposed, amounts.
10 · Methodology & sources
How this was built
Every figure here is reconstructed from public NYC Open Data (Socrata): DOB NOW filings, ECB and HPD violations, OATH case status, and HPD registrations. Firm identity is attributed from the DOB NOW filing-representative field and entity-resolved (messy free-text names normalized and conservatively merged). Penalty and collection figures are summed directly; the only estimated number is expediter revenue, which rests on a single stated fee assumption.
DOB NOW filing-representative attribution, ~2016–present (cleanest from ~2021). Aggregate figures only. This is a public, aggregate-only presentation — no individual firm-level target data or acquisition shortlist is published here.