Construction · Deep dive

PermitFlow

TurboTax for construction permits — an AI-plus-expediter platform that researches, prepares, files, and chases building permits across 7,000+ US jurisdictions for contractors, developers, and enterprise chains.

emerging

The question that decides it: PermitFlow monetizes the applicant-side pain of 7,000+ inconsistent permitting authorities with a hand-curated requirements database plus human expediters dressed in AI workflow. Does that layer convert to durable software margins before the complexity it sells against erodes from both ends — municipalities standardizing digital intake through Accela, OpenGov (Cox, $1.8B, 2024) and instant-permitting statutes like California's SB 379, and rivals like GreenLite attacking the actual bottleneck (plan review) rather than the paperwork — or do Glassdoor's 2024-25 reports of hire-and-fire operations churn reveal a tech-enabled expediting agency whose unit economics never escape the payroll?

My take ★★ Very interesting ★ Interesting − Maybe ✗ Not interesting

Snapshot

PermitFlow is the current capital leader in applicant-side construction permitting software: a workflow platform plus embedded expediting team that researches requirements, assembles applications, files them with the authority having jurisdiction (AHJ), and chases them to issuance across 7,000+ US jurisdictions (company figure, December 2025). Founded in 2021 and through Y Combinator’s W22 batch, it has raised $90.5M — a $31M Kleiner Perkins-led Series A in February 2024 and a $54M Accel-led Series B in December 2025 at a reported ~$500M valuation — and says it has powered more than $20B in construction value (December 2025). It matters because permitting is the most visible chokepoint in US construction, and because three different attack vectors — applicant software, private plan review, and government-side modernization — are converging on the same bottleneck from different sides. PermitFlow has bet on the applicant side; whether that is the durable position is the open question.

Founding story

Francis Thumpasery grew up outside Washington, DC, watching relatives in construction lose months to city permit counters — the founding anecdote he repeats in nearly every interview. Samuel Lam took a stranger route to the same place: Harvard CS and statistics, a marketing agency (NXTFactor) co-founded as a student, then software engineering at Uber Eats, where he built reporting and payments workflow tooling that had to conform to idiosyncratic per-company and per-municipality requirements. Permitting is that problem at maximum entropy: every one of roughly 20,000 US permitting authorities has its own forms, fee schedules, portal (or fax machine), and unwritten reviewer preferences.

The two founded PermitFlow in 2021 and entered Y Combinator’s Winter 2022 batch with the “TurboTax for construction permits” framing: the customer answers questions about the project once, and the system determines which permits are needed, generates the jurisdiction-specific application package, and files it. The honest early version, as with TurboTax’s storefront predecessors, was heavily human — an in-house expediting team using internal software — with automation layered in as volume taught the system each jurisdiction. The company was initially based in Milpitas/San Jose, California, and relocated its center of gravity to New York by the December 2025 Series B, closer to the enterprise developer and specialty-contractor customers driving revenue.

How it works

Mechanically, PermitFlow is four systems stacked on one asset — a proprietary, continuously maintained database of permit requirements, fees, forms, and timelines across 7,000+ AHJs (company figure, December 2025).

First, research. A customer scopes a project (new Lennar subdivision phase, an HVAC changeout, a retail buildout) and PermitFlow’s Research Agent queries the internal database and, where needed, the AHJ’s own portal to determine which permits apply, what each costs, what documents are required, and the realistic timeline — the answer a local expediter used to keep in their head. Second, preparation: the platform pulls project data from the customer’s CRM, contracts, and files, auto-populates the jurisdiction’s forms, and flags gaps or likely reviewer objections before submission. Third, submission: PermitFlow files with the AHJ through whatever channel the jurisdiction actually accepts — an Accela or OpenGov portal, email, or a human physically standing at a counter. This is the unglamorous, load-bearing detail: full national coverage requires people, and PermitFlow employs expediting staff alongside the software. Fourth, monitoring: the system watches portals for status changes and reviewer comments, notifies the customer, coordinates comment responses, and updates time estimates.

Since 2024 the company has repackaged the automation layer as named AI agents — for permit research, inspections scheduling, and contractor license management across jurisdictions (BusinessWire, December 2025). The agents are best understood as automating PermitFlow’s own back office first: every task the software absorbs converts services cost into gross margin. Integrations with Procore, Autodesk Construction Cloud, and ServiceTitan (announced 2023) embed permitting inside the tools contractors already run.

Product and business overview

The sellable product has four named components: Permit Management (the core workflow: research, preparation, submission, tracking, one dashboard across all projects and jurisdictions); Expediting Services (PermitFlow’s own team handles the filing end-to-end, with tiered premium levels including 24/7 monitoring and white-glove handling); AI Agents (research, inspections, and license management, the Series B expansion thesis); and Integrations (Procore, Autodesk, ServiceTitan embeds). Target customers cluster in three segments: national homebuilders and developers (Lennar, Brookfield Properties), commercial GCs and owners’ reps (Cushman & Wakefield, Wright Construction Group), and high-volume specialty trades — solar, HVAC, roofing — where a single company may file thousands of near-identical permits a year across hundreds of jurisdictions. That last segment is the volume engine; the first two are the contract-value engine.

Business model and pricing

Revenue is booked in two streams: SaaS subscriptions and per-permit service fees. Nothing is published on the pricing page — quotes are custom — but third-party software directories (Software Finder; TrustRadius pricing summaries, 2025-26) place enterprise plans at roughly $500–$5,000 per month scaled by permit volume and seats, with per-permit fees layered on top to capture activity peaks and smooth seasonality, plus a minimum monthly services fee that at least one Capterra reviewer (2025) called out as a con. Premium expediting tiers reportedly lifted average revenue per user about 22% year-over-year (third-party estimate, 2025). Pass-through municipal permit fees are handled through the platform but are not PermitFlow revenue. The mix question — how much of revenue is software versus human expediting — is the company’s most important undisclosed number.

Traction over time

Taking the claims at face value, 10x from a ~$4-5M 2023 base implies revenue in the $40-50M range at the December 2025 Series B — consistent with a ~$500M valuation at ~10-12x revenue, but unverified; no audited figure exists publicly. Named customers (Lennar, Brookfield Properties, Cushman & Wakefield, Red Tail, Wright Construction Group; 2024-25 press) skew large, which supports contract value but concentrates risk.

Market analysis

The narrow software market is modest: construction permit management software was about $1.43B globally in 2024, projected to $4.22B by 2033 at 12.7% CAGR (Growth Market Reports, 2024); the adjacent digital-permit-issuance market was estimated at $2.61B in 2025 growing ~15.7% (The Business Research Company, 2025). The real prize is the spend it replaces: fragmented permit-expediting services and internal permitting headcount across a $1.6T US construction market (2024). The structural tailwind is genuine and measurable — single-family authorization alone averaged ~1.5 months in 2023 (NAHB/Census), complex commercial permits run 6-12+ months, and a Washington-state builders’ study (2024) put the average 6.5-month permit delay at $31,375 in holding cost per home. Permitting reform is now bipartisan policy territory, which cuts both ways: political pressure creates budget for tools, but reforms that genuinely simplify intake (instant-permit statutes like California’s SB 379 for solar) shrink the complexity an intermediary monetizes.

Competitive intel

The competitor set is in frontmatter; the shape of the fight matters more. PermitFlow’s true incumbent is not software but the fragmented expediter industry — local consultants it beats on scale and visibility but not on single-city relationships. Its best-funded direct rival, GreenLite (~$86M raised through Sept 2025), is running a different and arguably deeper play: private third-party plan review, which removes the city queue rather than navigating it, and has landed Walgreens, O’Reilly, and TD Bank. Pulley ($4.4M, 2022) validated the thesis but lost the capital race. Symbium encodes codes as computation for building departments — the long-term deletion threat. And the government-side giants, Accela (Berkshire/Francisco Partners) and OpenGov (Cox, $1.8B, Feb 2024, permitting module Sept 2024), own the portals PermitFlow files into; every improvement they ship to intake UX erodes the applicant-side pain premium.

History and evolution

• 2021 — Founded by Thumpasery and Lam; Bay Area.
• Winter 2022 — Y Combinator W22 batch.
• May 2023 — $5.5M seed led by Initialized Capital.
• 2023 — Integrations shipped with ServiceTitan and Autodesk Construction Cloud; ~$4.1M revenue on 27 people (Getlatka).
• Feb 2024 — $31M Series A led by Kleiner Perkins; ~20 employees; national-expansion push begins.
• 2024-25 — Headcount scales past 100; Glassdoor reviews from this period describe rapid hire-and-fire cycles; HQ shifts to New York; Procore embedded integration.
• Sept 2025 — Rival GreenLite raises $49.5M Series B from Insight Partners — the category’s second big check.
• Dec 2, 2025 — $54M Series B led by Accel at reported ~$500M valuation; AI Agents for inspections and license management launched.
• Mar 2026 — 124 employees (Getlatka).

No pivots, no down rounds, no public crises — but also no disclosed revenue, and a scaling pattern (below) that suggests operational strain behind the clean press narrative.

What people say

The case for. Capterra reviews (2025-26) are strongly positive and specific: an assigned team that researches each municipality to confirm submittal details before filing, thorough time-stamped reporting, all communication and documents in one interface, and execution that one reviewer said fully exceeded expectations. Glassdoor (63 reviews, 2026) sits at 4.0/5 with 74% recommending — employees credit smart colleagues, a real mission, and strong sales culture. Investors effectively re-underwrote the story twice in 22 months: Kleiner Perkins led the A, and Accel led the B with every major prior investor participating (Dec 2025).

The complaints. Customer-side gripes are few but pointed: the minimum monthly services fee, and at least one Capterra reviewer (2025) reporting turnaround speed slowing as the company scaled — a warning sign for a business whose pitch is speed. The employee-side record is harsher. Recurring Glassdoor themes (2024-25 reviews, including ones titled “Brutal” and “Stay away from PermitFlow!”) describe burnout treated as lack of commitment, poor policy communication, high turnover, and a cycle of overestimating incoming work, hiring fast, then laying people off when volume disappoints. For a company whose delivery still depends on human expediting throughput, operations-team churn is not an HR footnote — it is a product-quality risk, and it is consistent with the customer report of slowing turnarounds. No G2 presence of note and no Reddit contractor threads of substance were found (searched July 2026), which itself says the specialty-trade long tail is still early.

Outlook: the open question

PermitFlow works if the AI agents genuinely convert expediting labor into software gross margin — visible as headcount growing far slower than claimed revenue (124 people on a claimed ~10x revenue jump by March 2026 is an encouraging early ratio) — and if the 7,000-jurisdiction database compounds into a data moat faster than intake standardizes. The bull case: permitting pain is politically salient and worsening in gateway markets, no rival matches its capital or jurisdiction coverage, the specialty-trade volume engine (solar, HVAC) creates TurboTax-like repeatability, and every permit filed teaches the system another reviewer’s quirks. The bear case: the moat is a phonebook — hand-maintained facts about jurisdictions that decay as Accela and OpenGov modernize the very portals PermitFlow scrapes and files into, while GreenLite removes the queue itself via private plan review and Symbium-style computational permitting deletes the complexity outright; meanwhile Glassdoor’s hire-and-fire pattern suggests margins still ride on services labor, making this a well-funded expediting agency at a ~$500M software price (Dec 2025). Watch three things through 2026-27: whether PermitFlow discloses (or leaks) a software-vs-services revenue split; whether headcount stays flat while revenue grows; and whether it follows GreenLite into third-party plan review — the tell that navigating the queue was never the endgame.

How a challenger would attack it

Attack the queue, not the paperwork. PermitFlow’s product navigates the city’s review line; GreenLite has already shown the deeper cut — private third-party plan review under statutes that remove the line entirely — and landed Walgreens, O’Reilly and TD Bank doing it. A challenger follows that wedge into PermitFlow’s core accounts: the national homebuilder or specialty chain doesn’t want a better-managed six-month wait, it wants a two-week approval, and only the review-side player can sell that. The second vector is the moat’s decay rate: PermitFlow’s asset is a hand-curated phonebook of 7,000+ jurisdictions, but the facts in it are public and increasingly machine-readable — a 2026-vintage challenger pointing frontier-model agents at AHJ portals rebuilds the research layer at a fraction of the cost PermitFlow paid, without the expediting payroll. Third, exploit the services drag: Glassdoor’s hire-and-fire pattern and a customer report of slowing turnarounds say delivery still rides on human throughput, and the minimum monthly services fee annoys exactly the high-volume specialty trades (solar, HVAC) that generate the repeatable permits. A pure-software product priced per permit with no minimum — TurboTax without the mandatory accountant — takes the volume engine while PermitFlow defends enterprise contracts with people it keeps churning through.

Same playbook, new buyer

Same database, opposite side of the counter — or a different counter entirely. The most valuable version of PermitFlow’s jurisdiction knowledge may not be sold to applicants at all: licensing the requirements-and-timelines dataset to lenders, title insurers and institutional developers as underwriting data (what will this project’s permit risk cost in months and dollars — the $31,375-per-home holding cost the Washington study quantified) monetizes the asset without any expediting labor, and PermitFlow’s applicant-side conflicts make it awkward for the incumbent to sell neutrality. The second shift is vertical: the same research-prepare-file-chase loop applies to liquor licenses, health permits, signage and fire inspections for multi-site retail and restaurant chains — buyers PermitFlow already reaches (Cushman & Wakefield-managed portfolios) but a regulatory domain its construction-shaped database doesn’t cover, and where no funded challenger exists. Third: utility interconnection queues for solar and EV infrastructure, where the specialty trades PermitFlow serves face a second, slower bureaucracy the platform ignores. The incumbent won’t follow fast: its Series B thesis, agent roadmap and 124-person org are committed to construction permits, and every new regulatory domain means rebuilding the phonebook from zero — the exact cost that made its first one expensive.

Sources and further reading

• PermitFlow Raises $54 Million to Solve Construction’s Biggest Bottlenecks With AI (BusinessWire, December 2, 2025)
• This AI Startup Is Cutting Through Construction Red Tape (Forbes, December 2, 2025 — ~$500M valuation, customer names)
• This YC alum just raised $31M to build the ‘TurboTax for construction permitting’ (TechCrunch, February 21, 2024)
• PermitFlow advances construction permitting automation with $5.5M in new capital (TechCrunch, May 8, 2023)
• PermitFlow Raises $54M to Cut Permit Approval Times from Months to Days (AlleyWatch, December 2025 — Thumpasery interview)
• How PermitFlow hit $4.1M revenue with a 27 person team in 2023 (Getlatka, accessed July 2026)
• PermitFlow Reviews (Glassdoor, accessed July 2026 — 63 reviews, 4.0/5)
• PermitFlow Reviews (Capterra, accessed July 2026)
• GreenLite Raises $49.5M Series B to Advance the Privatization of Construction Permitting (PR Newswire, September 2025)
• Cox Buys Majority Stake in OpenGov, Prompting $1.8B Valuation (GovTech, February 2024)
• Construction Permit Management Software Market Research Report (Growth Market Reports, 2024 — $1.43B 2024 TAM)