What changed
FACT: The FTC published record 2025 numbers — $3.5B reported lost to imposter scams (nearly 3x since 2020), the #1 fraud category, ~1 in 3 fraud reports (ftc.gov press release, id 5673). INFERENCE: cheap fast LLMs now make per-message scam classification cost fractions of a cent, so a free-tier funnel is economically viable.
Why now
FACT: the FTC release is fresh (June 2026), so fear and search intent around 'is this a scam' are elevated right now. HYPOTHESIS: that attention is a spike tied to a news cycle, not durable purchase intent.
Converging signals
Two signals: (1) a large, growing PAIN quantified by the FTC, and (2) a cheap/fast LLM capability. This is a pain × cheap-capability convergence, i.e. a discretionary quick-win shape, NOT a mandate/forced-buyer shape — no one is compelled to buy.
Customer pain
Real and widespread: ordinary people (and adult children of elderly parents) get 'bank/IRS/security alert' messages and have no fast, trusted way to judge legitimacy before acting. The FTC dollar figure proves the losses are real; it does NOT prove people will pay to prevent them.
Who pays
Proposed: anxious individuals and adult children, $4.99/mo or $0.99/deep-check by card. This is a low-WTP, high-churn discretionary consumer buyer — the weakest buyer class for this founder, who explicitly avoids consumer apps and ad-spend-heavy distribution.
Solved today
FACT-adjacent: carrier spam/scam-likely labels (built into iOS/Android and carriers), bank fraud-alert lines, free reverse-lookup and 'is this a scam' forums/subreddits, and general-purpose ChatGPT/Gemini/Copilot — into which a user can already paste the exact same message for free.
Why current solutions are bad
Built-in filters catch known bad numbers but don't explain novel social-engineering text; forums are slow. HOWEVER the gap is thin: a free general-purpose chatbot already does the paste-and-explain job competently, which is the core of this product.
Proposed product
A single-input web app: paste text/email or upload a screenshot → scam-likelihood score + named manipulation tactics + a safe next step, with a shareable link for elders. Thin LLM wrapper over a system prompt.
MVP version
Buildable in days: one input box, OCR for screenshots, one LLM call with a scam-taxonomy system prompt, a result card, Stripe for credits. Genuinely small scope — the problem is not build difficulty, it's defensibility and distribution.
30-day build
Ship free tool, seed a scam-tactic taxonomy, SEO/content around specific scam types ('is this USPS text a scam'), post in elder-care and personal-finance communities. Measure free→paid conversion — expect it to be very low.
60-day build
Add screenshot OCR, a family/guardian share flow, and email-forwarding intake. Test a 'monitor my parent's texts' angle for higher WTP. Watch for clones — this will appear within weeks.
90-day revenue plan
Realistic outcome is thin subscription revenue dominated by churn; the deep-check credit model is the more honest monetization. Getting to meaningful MRR requires paid acquisition, which the founder avoids and which this margin can't fund.
Distribution path
SEO + content + community posting is the only zero-cost channel, and it's crowded with high-authority sites (FTC, AARP, banks, carriers). Reaching the actual payer (worried adult children) at scale realistically needs ad spend — a disqualifier per founder profile.
Pricing hypothesis
$4.99/mo or $0.99/deep-check. Fine unit economics on inference, but the price is too low to fund CAC and too high to beat 'just ask free ChatGPT'.
Technical difficulty
Low. That is a liability here, not an asset — a weekend-cloneable AI wrapper with no data moat, no proprietary feed, and no switching cost.
Legal / regulatory risk
Moderate liability exposure: telling a user something is 'safe' that turns out to be a scam (or vice versa) invites blame; needs clear disclaimers. Not a licensure blocker, but a reputational/liability drag on a consumer safety tool.
Platform dependency
Depends on a third-party LLM API for its core function and on search engines for its only free distribution channel — both outside the founder's control.
Founder fit
LOW. This is a discretionary consumer app with low WTP and ad-dependent distribution — three things the founder explicitly avoids. It uses none of his edges (government portals, forced-filer mandates, public records, industrial ops). No mandate, no per-filing monetization, no forced buyer.
Breakout potential
Capped. The topic is hot but the product is undifferentiated; any breakout attracts instant clones and incumbents (banks, carriers, AARP, the FTC itself) who can bundle it free.
Final recommendation
PASS / KILL. Real pain, real news hook, but a low-WTP discretionary consumer wrapper with no moat and an ad-dependent path to the payer — a poor fit for this founder and likely to fail its own kill test against free chatbots and built-in filters. If pursued at all, reframe as a white-label 'scam-check' API/widget sold to banks, credit unions, senior-living operators, or elder-care software (a reachable B2B buyer with existing spend) rather than a direct-to-consumer subscription.
Next action
Do not build the D2C app. Spend two hours validating the B2B pivot instead: check whether credit unions / elder-care platforms already pay for scam-education or fraud-alert content, since that buyer — not the anxious consumer — is the only version of this with a defensible, reachable customer.