What changed
FACT (from the June 1, 2026 DOL final rule text): the Department of Labor published a combined final rule that (a) establishes a new, longer Form LM-2 Long Form for the largest labor organizations, (b) revises Form LM-2 for organizations at or above the $350,000 receipts threshold, (c) makes a parallel revision to Form LM-3, and (d) updates the reporting thresholds for Forms LM-3 and LM-4. FACT: on June 11, 2026 DOL published a correction because the Form LM-2 Long Form instructions were missing a page and both the LM-2 Long Form and LM-2 forms contained formatting errors. FACT: the rule applies prospectively under LMRDA section 208. HYPOTHESIS (not in the excerpt): the filing continues to be made through the OLMS Electronic Forms System (EFS) at dol.gov/olms within 90 days of fiscal year end — this is the long-standing LMRDA regime, but the excerpt does not state it.
Why now
Two independent timing facts converge. First, the form set itself changed — every affected union's prior-year working papers, spreadsheet templates, and accounting-system chart-of-accounts mappings are now mismatched to the schedules they must populate. Second, the June 11 correction is direct evidence that DOL's own instruction package was defective on release, which means the filer population is currently working from an instruction set that shifted underneath them within ten days. A form revision plus a threshold revision is precisely the moment when a filer who previously self-prepared discovers they cannot, and when a filer who was an LM-3 discovers they are now an LM-2 (or vice versa). This window is roughly one fiscal-year cycle wide. Once treasurers and their accountants have built new templates for the revised schedules, the pain subsides and the switching moment closes.
Converging signals
The convergence is structural rather than thematic: a federal rule (the June 1 final rule), a defined and enumerable filer class (every labor organization covered by the LMRDA — locals, intermediate bodies, and nationals), a mandatory annual submission with a statutory deadline, and a government portal with no platform owner who can deplatform a third-party preparer. A second, weaker signal in the input is the IRS proposed rulemaking on information-reporting dollar thresholds (April 17, 2026; hearing noticed July 2, 2026). That is a genuinely separate rule about a different filer population; it appears in the evidence set because it is semantically similar (a threshold change in an information-reporting regime), not because it interacts with the LM forms. I am explicitly declining to treat the IRS item as corroborating evidence for this opportunity. It is retrieval noise with a real underlying document.
Customer pain
HYPOTHESIS throughout this paragraph — the input contains no complaint threads, job postings, or forum evidence, and I will not invent any. The plausible pain: a local union's treasurer is typically an elected working member, not an accountant. The LM-2 requires itemized disbursement schedules with per-payee detail above a threshold, functional classification of every disbursement (representational activities, political activities, contributions, general overhead, union administration), and reconciliation to the union's books. The union's books live in QuickBooks or a spreadsheet with a chart of accounts that maps to nothing DOL asks for. The work is therefore a manual re-classification of a year of transactions into schedules that just changed shape. Missing the 90-day deadline exposes officers to LMRDA enforcement. Note the honest gap: pain of this kind is well-attested in the accounting trade generally, but the input gives me zero primary evidence of unions complaining about it, and I have scored accordingly.
Who pays
Three candidate buyers, in descending order of reachability. (1) The union itself — the treasurer or secretary-treasurer of a local or intermediate body, who signs the form and bears the liability. (2) The CPA firm or fund-accounting practice that already prepares LM-2s for a book of union clients — a firm doing thirty LM-2s a year is a single sale that carries thirty filings, and this is almost certainly the highest-value and fastest channel. (3) The union's outsourced bookkeeper. FACT from the mandate: the filer class is defined by statute and sorted by receipts, so the addressable buyer list is not a guess — it is enumerable from OLMS's own public disclosure database of prior-year filings, which lists every filer, its form tier, and its total receipts. That public file is the single most valuable asset in this opportunity: it is a complete, free, pre-qualified prospect list with revenue segmentation already attached.
Solved today
HYPOTHESIS (the excerpt does not describe current practice): larger unions engage specialist CPA firms — Calibre CPA Group, Withum (formerly Bond Beebe), and Novak Francella are the names that recur in the labor-fund audit space — who prepare the LM-2 alongside the annual audit, typically as a fixed-fee engagement. Mid-size locals use a local bookkeeper plus DOL's own free EFS software, which validates and transmits the form but does nothing to help you decide what belongs on which line. Small LM-3 and LM-4 filers self-prepare directly in EFS. The critical fact for competitive analysis: DOL gives away the filing transport layer for free. Nobody will pay for submission. They will pay for classification.
Why current solutions are bad
The free EFS tool is a form, not a preparer — it validates arithmetic and accepts a submission, but the operator still has to decide, transaction by transaction, whether a check to a hotel was representational activity or union administration. The CPA route solves this but at engagement pricing that is disproportionate for a local with $600,000 in receipts, and it is delivered once a year by a firm whose real product is the audit. Neither route survives a form revision gracefully: the free tool changed under the filer, and the CPA firm re-does its mapping manually across its whole book of clients. The wedge is the mapping layer that neither incumbent sells: a durable, auditable chart-of-accounts-to-schedule crosswalk that survives the revision and can be re-run.
Proposed product
A web tool that accepts a QuickBooks general-ledger export, Excel trial balance, or bank-transaction CSV; determines the correct form tier from total receipts against the revised LM-3/LM-4 thresholds and the $350,000 LM-2 threshold; applies a learned-plus-rules classifier that maps each account and each transaction to the correct revised schedule and functional category; itemizes payees above the disclosure threshold; runs DOL's validation rules before the filer ever opens EFS; and emits both a completed, reviewable form and a working-paper package showing every classification decision with its rationale, so an officer can defend the filing. The working papers are the product, not the form. Whether to submit programmatically into EFS or hand the filer a validated import file is an open technical question I would resolve in week one — the founder's FMCSA Training Provider Registry experience is directly transferable to answering it.
MVP version
Narrow hard: LM-2 filers only, QuickBooks Desktop and Online exports only, one fiscal-year-end cohort (December 31 year-ends, which must file by March 31). Build the account-to-schedule crosswalk by hand against the corrected June 11 instruction package. Ingest the OLMS public disclosure database to (a) train and sanity-check the classifier against thousands of real prior-year LM-2s, which is an unusually rich labeled dataset that exists because these filings are public by statute, and (b) generate the prospect list. Ship a single-tenant tool that a CPA firm can run for one client, produce the form and the working papers, and compare against last year's filed LM-2 line by line. That last comparison — 'here is what changed on your form because the rule changed' — is the demo that closes.
30-day build
Read the June 1 rule and the June 11 correction in full, including the corrected LM-2 Long Form instructions page that was missing. Build the crosswalk spreadsheet: every line of every revised schedule mapped to the account types that feed it. Pull the entire OLMS public disclosure dataset and reconstruct the filer universe — exact counts by form tier, receipts distribution, fiscal-year-end distribution, and the identity of the accounting firms that appear repeatedly as preparers. That last query converts a guessed market size into a measured one and hands you a ranked channel list. Do not write product code this month. If the OLMS data shows the LM-2 filer count is at the low end of my estimate and the preparer market is concentrated in four firms, kill the idea here for the cost of two weeks.
60-day build
Build the ingestion and classification engine against real prior-year filings, scoring the classifier by whether it reproduces the filing the union actually submitted. Recruit three design partners: one mid-size local that self-prepares, one CPA firm with a labor book, one intermediate body newly pushed across a threshold by the revision. Charge them. A free pilot with a union teaches you nothing about willingness to pay, and this founder's stated mode is selling through demonstrated value, which means the demo must be 'here is your own last filing, reproduced, and here is what the new rule does to it.'
90-day revenue plan
First revenue should come from the CPA channel, not from unions directly, because a firm makes a purchasing decision in weeks and a union local makes one at a monthly executive-board meeting. Target: three firms at a per-filing rate, covering perhaps forty filings. The seasonal reality constrains this hard — December fiscal-year-end filers are due March 31, so a build starting in July 2026 is aiming at the Q1 2027 filing season. The honest read is that first meaningful revenue lands 150-210 days out, at the edge of the founder's stated 30-180 day window, and the money is concentrated in one quarter. That seasonality is the single most under-appreciated risk in this opportunity and it does not appear anywhere in the input data.
Distribution path
The OLMS public disclosure database names every filer and, in the signature block, the preparer. That is a free, complete, legally-public prospect list with the incumbent's name attached to each row — an unusually strong cold-outreach asset. Direct outreach to the preparer firms first; direct outreach to unions whose receipts crossed a threshold in the revision second, because those filers are the ones whose obligation just changed and who know it. Content: a plain-English guide to what the June 1 rule changed, published against the correction notice, aimed at treasurers searching for exactly that. Avoid: labor-movement conferences, union endorsements, and anything requiring an international's blessing. Those are trust-cycle plays measured in years.
Pricing hypothesis
Per filing, tiered by form: LM-4 at $99, LM-3 at $299, LM-2 at $899, LM-2 Long Form at $2,500-$4,000. Firm licenses at a per-filing rate declining with volume. Annual recurrence is automatic because the obligation is annual. HYPOTHESIS: these anchors sit meaningfully below a CPA engagement fee for the same deliverable, which is the wedge, but I have no primary-source evidence of current CPA pricing in the input and this must be validated in the first thirty days by simply asking three firms.
Technical difficulty
Moderate and well-bounded. The genuinely hard part is not the software — it is the classification judgment, because the functional categories on the LM-2 (representational versus political versus administration) are contested, occasionally litigated, and not mechanically derivable from a transaction description. A classifier that is confidently wrong on the political-activity schedule creates liability for the officer who signs. The correct architecture is assistive: the tool proposes, itemizes its confidence, and forces human sign-off on every low-confidence line. The prior-year public filings give a large labeled corpus to calibrate against, which is a real and unusual advantage. Programmatic submission into EFS may or may not be feasible; if it is not, the product is still complete without it.
Legal / regulatory risk
Real but manageable and mostly reputational. The founder does not become a licensed or certified party — preparing an LM report is not a licensed activity the way tax preparation is regulated, so this is not the founder-must-get-licensed flavor of compliance burden. The exposure is that a materially misclassified filing harms a client. Mitigate structurally: the human signs, the tool never files without an affirmative review of every flagged line, and the working papers document every decision. Errors-and-omissions coverage is a cost of doing business here, not a blocker. Separately and non-trivially: labor-organization financial data is politically sensitive, and a vendor that aggregates it across many unions will be asked hard questions about who else sees it.
Platform dependency
None in the deplatforming sense. DOL is not a platform owner that can revoke access; a government reporting system cannot exclude a preparer whose customers are statutorily required to file. The real dependency is inverted and worth stating plainly: DOL itself ships free EFS software, and nothing prevents DOL from improving it. The rule-driven form revision is the moment DOL is most likely to invest in its own tooling. That is a live risk, but it bounds the product away from the transport layer and toward the classification layer, where DOL has no incentive to build — DOL does not want to tell a union what its disbursements mean.
Founder fit
Very high on shape, and this is not a close call. The founder has already shipped a production tool that reads a federal mandate, identifies the compelled filer class, builds against a federal portal (the FMCSA Training Provider Registry), and charges per submission. LM filing is the same object with a different agency: the rule compels, the class is enumerable, the portal exists, the fee is per filing, and the revenue recurs annually because the obligation recurs annually. Two fit caveats worth naming honestly. First, the founder's stated aversion to long trust-building plays runs directly into the culture of the labor movement, which is among the most vendor-skeptical buyer populations in the American economy — an outsider selling software to a union treasurer is starting from behind in a way an outsider selling to a trucking school is not. The CPA-firm channel routes around this and is the reason the idea survives. Second, the seasonality means this is not a 30-day cash business regardless of how fast the build goes.
Breakout potential
Bounded, and I want to be direct about it rather than optimistic. Using my own estimate of roughly 20,000-25,000 total LM filers with a few thousand at the LM-2 tier, and the pricing above, the entire filer population priced perfectly yields something in the low single-digit millions of dollars annually — and no one captures an entire compelled-filer market. A realistic ceiling for a solo operator with a good CPA channel is a few hundred thousand dollars of highly durable, highly renewing annual revenue. That is a genuinely good outcome for one person and a genuinely bad one for anyone expecting a venture-scale business, which is exactly the trade this founder says he wants. Note the honest weakness in that estimate: the filer count is my inference, and the OLMS public database would replace it with a measured figure inside a week. The real expansion path is horizontal — the same ingest-classify-validate-file engine points at any recurring financial disclosure filed by a small organization with no accounting staff, and Form 5500 for benefit plans is the obvious and much larger adjacent target. The LM work is arguably best understood as a cheap, low-competition proving ground for that engine rather than as the destination.
Final recommendation
Pursue, but as a thirty-day investigation that must earn the right to become a build — and be clear-eyed that the strongest version of this is a data play, not a filing play. The mandate shape is close to ideal for this founder and maps almost one-to-one onto the FMCSA product he has already shipped. What makes it unusual is that the compelled filings are public by statute, which hands you three things almost no compliance product gets: a complete prospect list, the incumbent preparer's name next to each prospect, and a large labeled training corpus of correct prior-year filings. Spend the first month in that OLMS dataset before writing a line of product code. It will answer the two questions that actually decide this — how many LM-2 filers there really are, and whether the preparer market is diffuse enough to be a market or concentrated enough to be four phone calls. If the preparer side turns out to be concentrated in a handful of firms, kill it immediately; that is not a software business and no amount of good execution makes it one. If it is diffuse, build for the CPA channel first and never sell directly to a local union, because the trust cycle there will consume the timeline. Understand also that the ceiling is low: a few hundred thousand dollars of durable recurring revenue for one person, which is a good outcome on its own terms and a poor one if you were hoping the market would grow into you. The real prize is the ingest-classify-validate engine, and the honest way to hold this opportunity is as a low-competition proving ground for that engine before pointing it at Form 5500, where the filer population is two orders of magnitude larger and the same architecture applies. Build LM-Filer to learn, price it to be profitable, and do not let it become the destination.
Next action
Download the complete OLMS public disclosure database of prior-year LM filings and answer three questions in one week: (1) the exact count of filers by form tier and the receipts distribution around the $350,000 threshold, which converts my inferred market size into a measured one; (2) the identity and filing volume of the preparer firms named in the signature blocks, ranked — if the top four firms account for most LM-2s, kill the idea on the spot; (3) the fiscal-year-end distribution, which tells you precisely how seasonal the revenue is and whether the Q1 2027 window is the only one that matters.