What changed
FACT: A reported wave of ADA disability-access lawsuits is hitting US cafe and shop owners (Guardian, signal cited). FACT: A new low-cost fast-tier Gemini model (3.7 Flash) makes the expensive part of an accessibility audit — converting raw WCAG violations into a prioritized, plain-English remediation narrative — nearly free to generate. HYPOTHESIS: These two together open a price point ($199) below both overlay subscriptions and agency audits.
Why now
The lawsuit wave is in the news cycle now, which is exactly when a demand-letter recipient or a scared neighbor of a sued business makes a panic purchase. Meanwhile the dominant cheap option — overlay widgets — is increasingly rejected by courts and even attracts suits itself, discrediting the incumbent answer at the moment demand spikes. This window is time-boxed: panic demand decays with news coverage.
Converging signals
(1) Complaint signal: Guardian coverage of a slew of ADA access lawsuits frustrating small shop owners. (2) Capability signal: Gemini 3.7 Flash makes per-report LLM cost cents, so a $199 one-time report carries software margins. (3) Structural signal (inference): ADA Title III liability is a standing legal obligation with no small-business exemption, so the underlying exposure recurs regardless of news cycle.
Customer pain
A shop owner opens a demand letter asking for $10k–$25k settlement plus remediation. They don't understand WCAG, their web guy doesn't either, a lawyer bills $300+/hr to tell them to 'fix the site,' and the $490/yr overlay they were sold may have made them a bigger target. They need, within days: what exactly is wrong, what to fix first, and a document to hand their lawyer showing good-faith remediation. CAVEAT (fact from source framing, not verified): the Guardian wave may be substantially about PHYSICAL premises access (ramps, counters, bathrooms), which a website scanner does not address — this must be validated before build.
Who pays
The sued or freshly-threatened owner of a small brick-and-mortar business with a website (cafe, restaurant, retail, salon) — a card-today discretionary buyer in acute distress. Secondary buyers: their defense lawyer (white-label reports per case) and small web agencies serving them (bulk audits). Serial ADA plaintiffs file hundreds of near-identical suits, so the defendant class is continuously replenished and publicly enumerable via dockets.
Solved today
(a) Overlay widgets — accessiBe/AudioEye/UserWay at ~$490+/yr — which courts and the accessibility community increasingly reject (per the convergence input); (b) $5k+ agency/law-firm remediation audits; (c) free scanners (WAVE, Lighthouse, axe DevTools) that dump raw technical violations the owner cannot interpret; (d) doing nothing and settling.
Why current solutions are bad
Overlays don't actually remediate and may increase legal exposure; agencies are 25x the price a cafe owner will pay and take weeks; free scanners produce output only a developer can use and carry zero legal-narrative value. Nobody sells the middle: a fast, honest, prioritized, plain-English report priced as a panic purchase.
Proposed product
A self-serve scanner: enter URL, pay $199, receive within the hour a branded PDF — axe-core WCAG violations, deduplicated, ranked by litigation relevance (the handful of failure types that actually appear in complaints: missing alt text, keyboard traps, unlabeled forms, contrast, missing captions), each with a screenshot, a plain-English explanation, and a copy-paste fix instruction for Squarespace/Wix/WordPress. Upsell: $499 done-for-you fixes on those three platforms; $29/mo re-scan monitoring for 'we fixed it and stayed fixed' evidence.
MVP version
Landing page + Stripe payment link + a Python worker: Playwright crawl (top 20 pages), axe-core per page, Gemini 3.7 Flash to write the narrative from a fixed rubric, WeasyPrint PDF, email delivery. All components are free/OSS except pennies of LLM and hosting. Genuinely a 1–2 week solo build; the founder has shipped harder.
30-day build
Week 1–2: build MVP; write the report rubric from 10 real ADA web-accessibility complaints pulled from PACER/CourtListener so the risk-ranking mirrors what plaintiffs actually plead. Week 2–4: run the kill test exactly as specified in the input — pull 30 recently sued businesses from public dockets (CourtListener federal ADA Title III filings; NY/FL/CA dominate), cold-email each a free 1-page teaser scan of THEIR site with 2 real violations shown, CTA to the $199 full report. Kill threshold: <3 conversions in 2 weeks.
60-day build
If ≥3 convert: automate the docket→teaser pipeline (daily CourtListener pull, auto-scan, auto-personalized outreach — this is complaint-mining plus public records, squarely the founder's strengths). Add the $499 done-for-you tier using a contractor for fix execution. Approach 5 ADA defense firms with white-label per-case pricing ($99/report at volume).
90-day revenue plan
Target: 30–50 report sales ($6k–$10k) + 10 done-for-you ($5k) + 1–2 law-firm white-label accounts. Recurring layer: monitoring subscriptions converting ~20% of report buyers. If docket outreach converts near 10%, the channel alone sustains this; if it converts near 1%, the business is a trickle and should be folded into a broader compliance-report product or killed.
Distribution path
Primary: outbound to named, publicly-identified defendants from court dockets — a finite, high-intent, continuously refreshed list nobody else is systematically working at this price point. Secondary: SEO for 'ADA website demand letter what to do' (low competition, exact panic query); defense-lawyer referral; Squarespace/Wix web-designer communities. Weakness: outbound to sued businesses can read as ambulance-chasing and email deliverability/response from stressed owners is unproven — this is the single biggest risk.
Pricing hypothesis
$199 one-time report (panic-purchase price, no procurement); $499 done-for-you fixes; $29/mo monitoring; $99/report white-label at 10+/mo for law firms. All card-today, no contracts.
Technical difficulty
Low. axe-core, Playwright, an LLM prompt, and PDF generation are commodity parts; the moat, such as it is, is the litigation-informed rubric and the docket-outreach machine, not the scanner.
Legal / regulatory risk
Moderate and manageable: the report must be framed as a technical accessibility assessment, NOT legal advice or a compliance guarantee — no 'this makes you ADA compliant' claims (automated scanning catches only ~30–40% of WCAG issues; the report must say so honestly, which is also the honest-wedge positioning). Unauthorized-practice-of-law risk is low if positioned as technical findings for the owner and their counsel. Solicitation of litigants is legal for a non-lawyer service but must not imply legal representation.
Platform dependency
None meaningful. Gemini is swappable for any cheap LLM tier; axe-core is OSS; no marketplace approval, no platform owner who can deplatform.
Founder fit
Good but not the core thesis. This is not a government-portal forced-filer play — no agency, no filing, no per-submission monetization into a portal. It IS: complaint-mining, public-records mining (dockets), AI-workflow automation, low-budget fast prototyping, and demonstrated-value selling (the free teaser scan of the prospect's own site is exactly his sales motion). The lesson favoring mandate plays (conf 0.65) applies and caps fit below the 9s; the quick-win rubric applies fully — this is a days-to-weeks build with card-today buyers.
Breakout potential
Moderate. The same docket→scan→teaser→report machine generalizes to other lawsuit-driven compliance panics (BIPA, CIPA/wiretapping web suits, NY SHIELD). The white-label law-firm channel could turn it into a per-case litigation-support data product — a durable, less-copyable business than the consumer-facing report.
Final recommendation
PURSUE THE KILL TEST ONLY — do not build past the MVP until it passes. This is a legitimate quick-win-shaped opportunity: acute pain, discredited incumbents, card-today buyer, 2-week solo build, and a distribution idea (docket mining) that matches the founder's public-records strength. But it rests on one news article, an unverified assumption that the lawsuits are web-based, and an unproven outreach channel. The $199 test costs under two weeks and ~$100. Run it; the data decides.
Next action
Before writing code: pull the last 60 days of ADA Title III federal filings for NY/FL/CA from CourtListener (free API), read 15 complaints, and count what fraction allege WEBSITE inaccessibility vs premises. If ≥1/3 are web-based, build the MVP and run the 30-defendant teaser test exactly as specified in the input.