What changed
FACT: On 2026-06-17 FDA issued a final rule classifying 'radiological machine learning-based quantitative imaging software with predetermined change control plan' into class II with special controls (Federal Register 2026-12166). The special controls are codified and condition the entitlement to ship model updates without a new 510(k) on operating under, and documenting conformance with, a pre-approved PCCP.
Why now
FACT: the classification is newly codified, so the first cohort of devices cleared under it will hit their first post-clearance model-update cycles over the next 6-18 months. HYPOTHESIS: that cohort is managing PCCP conformance in Word/spreadsheets because eQMS incumbents price above seed-stage budgets and their PCCP-specific modules are immature. This hypothesis is untested — it is the core validation task before building.
Converging signals
One strong regulatory signal (the class II + PCCP special-controls rule) converging with the known structural pattern of gated-entitlement paperwork. NOTE: the signals array is empty and only one demand_evidence item was supplied, so convergence breadth is thin — this is closer to a single-signal hypothesis than a multi-signal convergence.
Customer pain
HYPOTHESIS: RA/QA leads at 5-30 person AI device startups must execute every model update exactly per the cleared PCCP — verification protocols, acceptance criteria, signed records — forever, and fear an FDA inspection finding of nonconformance that voids their no-new-510(k) entitlement. No PAIN-type evidence (complaints, forum posts) was provided, so felt pain is unproven; the obligation is proven, the suffering is not.
Who pays
Regulatory affairs / quality leads at seed-to-Series-A AI-imaging and digital-health device companies that cleared (or are drafting) a PCCP; secondarily, medtech regulatory consultants who author PCCPs for such clients and would resell templates/workflow.
Solved today
HYPOTHESIS: Word/Excel change records inside a general eQMS (Greenlight Guru, Qualio) or no eQMS at all pre-clearance; consultants author the PCCP as a static document with no execution tooling. Ketryx already sells validated CI/CD + change control aimed specifically at software/AI medical devices, which partially solves this for funded companies.
Why current solutions are bad
Manual records drift from the cleared plan; evidence capture from ML pipelines is copy-paste; nothing enforces that each update followed the exact pre-approved protocol, which is precisely what an FDA inspector will check. But 'bad' is only monetizable if inspections/audits are feared and update cadence is monthly-to-quarterly — both unverified.
Proposed product
(1) Wedge: PCCP + special-controls authoring template pack ($500-2,000 one-time) matched to the new classification regulations, sold to startups and consultants. (2) Core: SaaS that ingests the cleared PCCP, generates the per-update checklist, pulls validation evidence from CI/ML pipelines via API/webhook, produces signed, immutable change records and a one-click auditor export. Priced $300-800/mo per company.
MVP version
Template pack + a thin web app: structured PCCP editor, update-run checklist with e-sign, file/metric evidence attachment, PDF/CSV audit export. Skip deep CI integrations until a paying pilot asks.
30-day build
Validation sprint, not code: pull openFDA/510(k) database counts of clearances citing PCCPs to size the live cohort; interview 10 RA/ML leads at AI-imaging startups (target per the testable prediction: ≥6 on spreadsheets, ≥3 willing to pilot at $300-800/mo); interview 3 medtech regulatory consultants on whether they'd resell templates. Kill or proceed on these numbers.
60-day build
If validated: ship the template pack and sell it via consultants and direct LinkedIn outreach to RA leads at recently cleared companies (clearance holders are public in the 510(k) database — a reachable, enumerable list). Build the checklist/e-sign/export MVP with 2-3 design partners.
90-day revenue plan
Template-pack sales ($500-2,000 each) plus 2-3 paid pilots at $300-500/mo. Realistic first-revenue path is the templates at day 60-90; SaaS revenue by day 120-180 given medtech vendor-qualification friction.
Distribution path
Enumerable buyer list from FDA's public 510(k) clearance database (companies citing PCCPs), direct outreach to named RA/QA leads, partnerships with regulatory consultants who author PCCPs, content in RAPS/medtech-regulatory communities. Narrow but precisely targetable; the constraint is small N, not findability.
Pricing hypothesis
$500-2,000 one-time templates; $300-800/mo per company for the workflow SaaS (deliberately under eQMS price floors). Per-update pricing is tempting but cadence is unknown — a falsification risk if updates are annual.
Technical difficulty
Low-moderate for the MVP (forms, e-sign, immutable records, exports). Moderate for CI/ML pipeline evidence capture. The hard part is not code: it is that software used within a customer's quality system typically must be validated by the customer, and buyers will ask for the vendor's own quality posture (SOC2-ish expectations, validation docs).
Legal / regulatory risk
The product itself is not a medical device and needs no FDA clearance (it is recordkeeping tooling). Risk is commercial-regulatory: if the tool generates or holds QMS records, customers must qualify it as a supplier, adding sales friction rather than legal exposure.
Platform dependency
None material — no app-store or marketplace gatekeeper. Dependency is on FDA policy continuing to expand PCCP pathways (trend currently favorable per the cited rule).
Founder fit
Mixed, and weaker than surface pattern-matching suggests. The gated-entitlement-paperwork shape matches his proven ELDT edge, BUT the ELDT win had a government PORTAL and per-filing monetization; PCCP records are kept internally and produced at inspection — there is no portal to automate and no per-transaction billable event. He also has no medtech/regulatory-affairs credibility, and this buyer class vets vendors on exactly that. His systems-thinking and AI-workflow strengths apply to the product; his distribution edge does not. The accumulated lesson scoring portal-mandate opportunities 8-9 does NOT apply here because there is no portal submission.
Breakout potential
Moderate: if PCCP becomes the default for AI device updates across categories (FDA signaled this direction), the tool generalizes to every AI-SaMD company, and adjacent expansion exists (EU MDR change control, EU AI Act technical documentation). Ceiling capped by incumbents (Ketryx, eQMS vendors) adding PCCP modules once the market proves out.
Final recommendation
DO NOT BUILD YET — run a 2-3 week validation sprint first. The mandate is real and newly codified, the buyer is enumerable from public FDA data, and the wedge (templates via consultants) is cheap to test, so this clears the kill bar for a probe. But cohort size, update cadence, felt pain, and incumbent coverage (especially Ketryx) are all unverified, and founder fit is materially weaker than his portal-filing pattern. Proceed only if the interview and openFDA counts hit the stated thresholds.
Next action
Query openFDA's 510(k) endpoint for clearances referencing predetermined change control plans to count the live cohort, and simultaneously message 10 RA/QA leads at AI-imaging startups (sourced from recent clearance records) asking how they document PCCP conformance today.