AllergenTruth Clear

Executive summary

AllergenTruth Clear is a done-for-you allergen-declaration truth service for U.S. FDA-regulated packaged-food brands and co-packers. The customer does not operate software; they receive a SKU-level Allergen Truth Binder that reconciles the bill of materials, supplier allergen statements, and draft label artwork against FALCPA/FASTER Big-9 requirements, with a print-go / print-hold decision memo released by a human food-label specialist.

Beachhead: independent specialty CPG brands and co-packers with 5–80 SKUs and no dedicated regulatory labeling FTE. Primary offer is outcome-priced (per SKU / portfolio / continuity desk), never hourly. AI extracts and crosswalks; humans release. The company never claims to practice law or replace the facility’s PCQI / label owner of record.

Thesis

Undeclared allergens remain the stubborn #1 structural failure mode in U.S. packaged food — not because the Big-9 list is unknown, but because formula changes, supplier sub-ingredient swaps, artwork freezes, and Contains-statement logic live in disconnected tools. A narrow AI-native service that sells print-release-ready allergen truth (not a co-pilot, not a full NFP generator) can capture existing label-consultant and QA budget with higher throughput and a change-control continuity wedge.

Discovery rationale

This run deliberately steered away from the saturated “regulatory filing completeness engine” pattern and from already-claimed self-storage lien, ocean D&D, LOTO, PIT certification, lodging-tax remittance, and restaurant/cottage allergen desks. Fresh searches across logistics, hospitality tax, LOTO, plasma, MSB, carnival permits, swim/CPO, and food labeling showed the strongest CODE + budget + remote-fulfillment fit in commercial packaged-food allergen declaration truth — adjacent to but hard-differentiated from GhostTrue (virtual kitchen menus), CottageLabel (CA cottage food county packs), CA-ADDE menu packs, DrillTrue (mock recall drills), and supplement claim clearance.

Candidate comparison

CODE validation

Rubric scorecard (selected market)

Six-gate average: 4.67. Evidence threshold: cleared (buyer, pain, spend, demand, competitors, MVP, margin path, licensing boundary).

Target buyer

• ICP: U.S. FDA-regulated packaged-food brand or co-packer, 5–80 SKUs, $2M–$80M revenue, no full-time labeling regulatory specialist.
• Economic buyer: Founder/CEO, VP Quality/Food Safety, or Head of R&D/Regulatory.
• Champion: QA manager, PCQI, or brand ops lead who owns label artwork freeze.
• Trigger: New SKU launch, formula change, supplier change, co-packer onboarding, retailer audit ask, near-miss Contains mismatch, or post-warning-letter remediation.

Jobs-to-be-Done

• When I freeze artwork, help me know every Big-9 allergen is correctly declared before I spend on print.
• When a supplier updates a spec, show me which SKUs break Contains/ingredient truth.
• When my retailer or auditor asks “prove your allergen declaration,” give me a defensible binder in hours, not weeks.
• When I change a formula, tell me go/hold without me operating labeling software.

Painful problem

Undeclared allergens drove ~34% of FDA food recalls in 2025 and 57 FDA events in Q1 2026 alone. Label errors (mostly undeclared allergens) were estimated to cost the industry ~$1.92B in 2024 direct recall expenses (Loftware/FDA Enforcement Report analysis using ~$10M average direct cost). Small brands still reconcile Excel BOMs, PDF supplier sheets, and Illustrator proofs manually — and miss sesame/sub-ingredients/Contains completeness.

The outcome we sell

Print-release-ready allergen declaration truth for each SKU — a specialist-released Allergen Truth Binder stating go or hold, with a Big-9 matrix, Contains/ingredient reconciliation, defect list, and evidence citations. Not a dashboard. Not a co-pilot. Not legal advice.

First one-feature MVP wedge

Evidence summary

• Undeclared allergens #1 FDA recall cause (Food Safety Magazine 2025 analysis; Sedgwick Q1 2026).
• FASTER Act sesame + FDA Edition 5 allergen Q&A (Jan 2025) keep declaration rules operationally non-trivial.
• Existing paid alternatives: Food Lab ($480–$840 kits), consultant reviews ($200–$500+), lawyer reviews (~$500+), SaaS LabelCalc/ReciPal, enterprise Genesis/TraceGains.
• ~91k domestic FDA-registered food facilities (Mar 2024); ~22k food manufacturing companies.
• Warning letters (e.g., One Roof LLC May 2025; Revolution Foods Apr 2025) cite allergen preventive controls and misbranding.

Claim table

Source-claim matrix

Market and demand evidence

Packaged food allergen misbranding is a recurring, budgeted risk. Loftware’s analysis of FDA Enforcement Report data attributed nearly half of 2024 food recalls to label errors, overwhelmingly undeclared allergens, with an estimated $1.92B industry direct cost using the ~$10M average recall benchmark. Sedgwick’s Q1 2026 index shows undeclared allergens still leading FDA food events (57). Buyers already buy SaaS (LabelCalc/ReciPal), service kits (Food Lab), enterprise stacks (Genesis/TraceGains), and counsel. The service wedge is the release-ready truth pack for teams that will not staff a labeling regulatory desk.

Active buyer conversations

• FDA Warning Letters and recall press cycles create public, time-bound buyer urgency.
• Trade press (Food Safety Magazine, National Provisioner, QA Magazine) repeatedly frames undeclared allergens as the perennial #1 problem.
• Vendor pages (Food Lab, Matrix Sciences, Mérieux, Trustwell, Food Label Maker “hire a consultant”) show buyers shopping managed review — not only DIY tools.
• LinkedIn/consultant webinars discuss lawyer label-review pricing starting ~$500/SKU, confirming willingness to pay for expert release.

Competitive landscape

Competitor and budget validation

Budget already exists in (1) SaaS subscriptions $29–$79+/mo, (2) per-label consulting $200–$840, (3) legal review $500+, (4) enterprise licenses $5k–$10k+/yr, (5) internal QA hours before every artwork freeze. AllergenTruth redirects those budgets into a DFY outcome with specialist release and a continuity retainer for supplier/formula change monitoring. Winning thesis: faster cycle + change-control flywheel + audit-ready binder, not “cheaper Genesis.”

Pricing evidence and proposed pricing

Never hourly. Never contingency on avoided recall value (creates misaligned liability optics). Rush 24h: +40%.

Regulatory and compliance considerations

• FALCPA (2004) + FASTER Act (sesame, eff. 2023) define major food allergens (Big-9).
• FD&C Act §403(w) misbranding if major allergens not declared; Contains statement rules matter.
• FSMA PC rule 21 CFR 117.135(c)(2): allergen preventive controls include labeling controls to prevent misbranding.
• FDA Guidance Edition 5 (Jan 2025) is the working interpretation layer for labeling edge cases.
• Service produces completeness/reconciliation packs; facility remains responsible for final label and PCQI decisions.

Licensing boundary

AI-native advantage

AI changes unit economics by collapsing multi-hour PDF/spec/artwork reconciliation into minutes, enabling specialist-only review. As models improve OCR, table extraction, and Contains-logic checking, COGS per SKU falls while the continuity desk scales across supplier-change monitoring — without converting the product into a customer-operated co-pilot.

Internal AI engine architecture (10 layers)

1. Intake — secure upload of BOM, supplier sheets, artwork PDF; SKU metadata; change-trigger type.
2. Normalization — OCR/parse ingredients, sub-ingredients, allergen statements, artwork panels.
3. Retrieval/knowledge — FALCPA/FASTER Big-9 rules, sesame edge cases, Contains completeness rules, gold examples.
4. AI workbench — draft allergen matrix; map sources → declarations; propose Contains statement.
5. Deterministic rules — FATAL if major allergen in formula missing from label; MAJOR if Contains incomplete when used; MINOR formatting.
6. Human chokepoint — specialist release / hold with rationale.
7. QA — second-pass on FATAL/MAJOR; red-team random sample.
8. Delivery — Truth Binder PDF + go/hold memo + defect CSV.
9. Learning loop — defect patterns → SOP/prompt/rule updates.
10. Model-portability — prompts/rules/eval sets portable across frontier models; never hard-lock to one vendor.

AI-vs-human operations pipeline

Dynasty translation layer

• Buyer: QA/founder pays to avoid Class I allergen misbranding before print.
• Service: DFY allergen truth pack + optional continuity desk.
• Workflow: Intake → extract → rules → specialist → QA → binder → change monitoring.
• Tooling: Drive/Dropbox intake, LLM OCR, rule engine, Notion/Airtable case system, Docx/PDF templates — software later.
• Sales: “We’ll tell you go/hold on allergen declarations before you print — free Gap Scan on 1 SKU.”
• Delivery: Manual-assisted first 5 pilots; harden rules after 10; pause scale until COGS measured at 20.
• Expansion: Co-packer onboarding packs → supplier change watch → retailer evidence portals → partner NFP labs.

Anti-duplication analysis

Checked manifest (908 runs restored from healthy tip 6cc6db99) and root *-blueprint.html filenames. Near-neighbors exist but differ in buyer+workflow+outcome:

• GhostTrue — virtual kitchen DBA + delivery menu allergen truth (restaurant ICP).
• CottageLabel — CA cottage-food county label submission.
• CA-ADDE menu pack — chain restaurant menu disclosure.
• DrillTrue — GFSI mock-recall drills, not label declaration.
• Supplement label-claim engine — DSHEA claims for supplements, not FALCPA packaged food.

AllergenTruth’s unique wedge: FDA packaged-food print-release allergen declaration truth with change-control continuity.

Anti-commoditization analysis

If frontier models make DIY extraction cheap, AllergenTruth still wins on (1) specialist release liability interface, (2) deterministic rule corpus + gold examples, (3) supplier-change monitoring ops, (4) audit-ready binders retailers accept, (5) measured rework/hold rates. The product is the operating system for allergen truth — not a chat window.

Service delivery workflow

1. Sales: Gap Scan → Truth Pack quote.
2. Kickoff checklist: BOM version, supplier statement dates, artwork PDF, prior Contains language.
3. AI draft matrix + defects within 24h.
4. Specialist review same day or next.
5. Client Q&A on holds (async).
6. Final binder delivered; optional Continuity Desk enrolls SKUs for change watch.

Operations as product

• Structured intake checklists; required evidence lists; completeness gates.
• Exception queues (missing supplier sheet, illegible artwork, dual FDA/USDA).
• Confidence scoring; reviewer assignment; audit trail; version control.
• Gold-standard binders; red-team undeclared-allergen tests; RCA on any post-release defect.

No-holes quality engine

FATAL defects block release. Two-person rule on any FATAL cleared as false positive. Monthly red-team: inject known undeclared sesame/milk cases into eval set. Customer-facing binder always shows evidence links to source pages/lines.

What the human expert actually does

Minimum viable offer

Allergen Truth Pack (1–5 SKUs): Free Gap Scan on SKU #1 → paid pack $349–$799/SKU → binder in ≤3 business days. Continuity Desk offered after first paid pack. No platform login required for the customer.

Fulfillment process (first 3 customers)

1. Shared Drive folder + intake form.
2. Founder/specialist runs AI draft in Claude/GPT + spreadsheet matrix template.
3. Manual Contains check against FDA Edition 5 notes.
4. Deliver branded PDF binder; Loom walkthrough of holds.
5. Automate only after 5 clean packs: OCR pipeline, rule engine, templated PDF.

Tools and systems (day one)

• Google Drive / Dropbox; Typeform/Tally intake; Airtable cases; Notion SOPs.
• Frontier LLM + PDF OCR; spreadsheet Big-9 matrix; Docx/PDF export.
• Stripe invoicing; Calendly for scan reviews; Gmail/CRM for outbound.
• Later: custom extractors, supplier portal, retailer export API — not required for revenue.

Human-in-the-loop quality control

No binder ships without specialist RELEASE. FATAL findings require explicit disposition. Random 10% dual-review. Any customer-reported post-release defect triggers RCA within 48h and rule update.

Nonlinear scaling and unit economics

Revenue/FTE target: $250k–$400k by month 12 with Continuity Desk mix. CAC payback target <60 days via Gap Scan → Pack. Waitlist→pilot 15–25%; pilot→paid 50–70%; retention on Desk 80%+ annual if change alerts prove value.

Distribution proof table

Sales and outreach plan

Lead with diagnosis, not demo: “Send one SKU — we’ll return a Gap Scan showing undeclared/Contains risks before you print.” Convert to Truth Pack same week. Continuity Desk pitched after first hold caught.

Founder-led content plan

Teach Big-9 edge cases (sesame, fish species, tree-nut specificity), Contains-statement traps, supplier sub-ingredient failures, and warning-letter teardowns. No generic “AI for food” content.

First 30 days of content

1. 10 educational posts: sesame FASTER traps; Contains incompleteness; co-packer BOM vs brand artwork; milk derivatives; spice blends; “may contain” ≠ control; artwork freeze checklist; retailer audit asks; warning-letter lessons; SKU change control.
2. 3 diagnostic teardowns: anonymized Gap Scan redacts; Contains mismatch before/after; supplier sheet vs label matrix.
3. 2 lead magnets: Allergen Artwork Freeze Checklist; Big-9 Sub-Ingredient Cheat Sheet.
4. 1 webinar: Live 1-SKU Gap Scan clinic.
5. 1 outbound diagnosis template: “I reviewed your public label PDF for Contains completeness — 3 questions before you print.”

Lead magnet and waitlist plan

Free Allergen Gap Scan (1 SKU) + waitlist for Continuity Desk. Buyer receives a 1-page risk memo. Pain signal captured: # FATAL/MAJOR defects. Sales-ready if ≥1 FATAL or artwork freeze <30 days. Follow-up within 24h with Truth Pack quote.

Warm GTM plan

Scan alumni, co-packer intros, food-accelerator cohorts, prior consultant clients, and specialty-food Slack/communities. Offer founding Portfolio Pack pricing for first 10 brands.

Targeted outbound plan

Build list of brands with recent new SKUs / retailer expansions. Personalize using public label PDFs. CTA = Gap Scan, not meeting. Sequence: day 0 PDF note → day 3 teardown snippet → day 7 case study → break-up.

Answer-engine / search visibility plan

Publish citation-backed pages answering: “Does Contains must list all major allergens?”, “Sesame labeling examples”, “Supplier change allergen checklist”. Structure for AI-overview extraction with primary-source links (FDA, FR, statutes).

Pilot design and early-demand-trap mitigation

• Pilot cap: 8 brands / ≤40 SKUs total.
• Incentive: founding Pack pricing (−20%) for published anonymized learnings.
• Feedback: defect false-positive/negative tags — not custom feature builds.
• Mitigation: refuse full NFP redesign and legal opinion work in pilot; partner out.

Early-access feedback flywheel

Every hold → root-cause tag (supplier / artwork / BOM / rule gap) → weekly SOP/prompt update. Product feedback = reusable rule. Custom work = one-off counsel escalation, not roadmap.

Build-before-scale checkpoints

• After 5 pilots: harden intake + evidence requirements + FATAL taxonomy.
• After 10: harden SOPs, exception queues, reviewer checklists, binder templates.
• After 20: pause new logos until COGS, rework, escalation, cycle time measured; no scaling by adding unchecked reviewer headcount.

7-day / 30-day / 90-day launch plans

7 days

Landing page + Gap Scan form; matrix template; 20 outbound; 3 co-packer intros; first scan delivered.

30 days

10 educational posts; webinar; 8–12 scans; 3 paid packs; first Continuity Desk LOI; rule eval set v1.

90 days

25+ paid SKUs; 5 Desk subscribers; COGS dashboard; partner counsel/lab shortlist; decide beachhead expansion (sauces → snacks → frozen).

Metrics and KPIs

• Scan→Pack conversion; cycle time; FATAL catch rate; false-positive rate; rework %; Desk logo retention; revenue/FTE; gross margin; NPS of binder clarity.

Risks and mitigations

See exhaustive register below. Top risks: scope creep into full label law; liability optics; enterprise incumbent response; thin differentiation if positioned as “AI LabelCalc.”

Exhaustive risk register

What could kill this

If buyers insist on end-to-end legal certification; if a dominant PMS/label SaaS bundles free specialist-quality allergen release; if we cannot keep false-negative rate near zero; if we broaden into generic AI consulting.

Go/no-go reasoning

GO. Clears evidence threshold: specific buyer, painful Class I problem, verified demand/spend, competitor budget, narrow MVP, remote fulfillment, licensing boundary, 50%+ margin path, hard-diff from prior allergen-adjacent blueprints. Rejected weaker adjacent candidates rather than forcing a carnival/MSB clone.

Final recommendation

Launch AllergenTruth Clear with a 1-SKU Gap Scan → Truth Pack wedge in specialty sauces/snacks/bakery, cap pilots at 8 brands, harden rules before scale, and sell Continuity Desk as the compounding revenue layer. Keep counsel/NFP partners at the edge. Do not build a customer-operated labeling platform first.

Source list

1. Food Safety Magazine — Recall Litigation and Regulatory Trends
2. Food Safety Magazine — Column-Reg Report June/July 2026
3. QA & Food Safety — Sedgwick Q1 2026 recall index
4. National Provisioner — 2025 recalls year in review
5. New Food — Loftware label-error recall analysis
6. TraceGains — Undeclared allergens risk
7. FASTER Act — Pub. L. 117–11
8. Federal Register — FDA Allergen Q&A Edition 5 notice
9. FDA — One Roof LLC Warning Letter
10. FDA — Revolution Foods Warning Letter
11. Food Lab — Services & Pricing
12. OrbitalJump — Nutrition labeling software cost comparison 2026
13. Matrix Sciences — Product Label Review
14. Mérieux NutriSciences — Food Label Compliance
15. Trustwell — Food Labeling Consulting
16. CIRS — FDA facility registration statistics
17. Kentley Insights — US Food Manufacturing Market Report 2025
18. Esko — Food Allergen Labeling Requirements 2026
19. The FDA Expert — Undeclared allergens prevention
20. Food Label Maker — Hire a labeling consultant

Generated 2026-07-26T18:05:00Z. Research ledger maintained during run; 18 targeted web searches across logistics, lodging tax, LOTO, plasma, MSB, carnival, swim/CPO, and food allergen labeling.