A mobile photo-capture app that produces cryptographically tamper-evident audit photos: every photo is hashed and signed at capture time, the captured bytes carry a visible timestamp + GPS overlay, and anyone can verify the photo is unaltered via a QR code or a public viewer URL.

In one sentence: "Photos that survive cross-examination."

The technical guarantee is that a verifier — an insurance adjuster, a court, an auditor — can prove a photo was taken at a given time and place by a specific app installation, without trusting either the photographer or our servers post-capture.

2. Why now

Generative AI has made image fabrication cheap. Insurance, building inspection, and claims handling are seeing first-wave fraud cases where AI-generated "damage" photos slip into legitimate workflows. The pain is acute and the buyer is willing.
EU eIDAS 2.0 (in force 2026) creates a regulated category of "qualified electronic attestation of attributes" that maps onto signed audit photos. Early movers get to define the practice.
NIS2 + GDPR Art. 30 push processors of evidence-grade media to document chain-of-custody, which a free iPhone Camera workflow can't satisfy.

3. Target customers

In rough order of accessibility:

| Segment | Pain | Typical photos / month / user | Willingness to pay | |---------|------|--------------------------------|---------------------| | Independent insurance loss adjusters | Damaged-property claims require court-defensible photos. Current workflow is photo + Word document — easy to dispute. | 100–400 | High — already paying for SaaS like Xactimate. | | Building inspectors (NL: bouwinspecteurs) | Statutory inspections of construction sites require photo-evidence reports per Wkb (Wet kwaliteitsborging). | 200–800 | High — billable to the construction project. | | Forensic auditors / fraud investigators | Evidence chain integrity is the job. | 50–200 | Highest, smallest volume. | | Real-estate agents | Listing photos liability if "before/after" disputes arise. | 30–100 | Medium — price-sensitive. | | Field-service technicians (HVAC, telecoms, utilities) | "Did the technician really do the work?" disputes between contractor and customer. | 50–300 | Medium. |

NL/BE first. EU within 12 months. UK/US after, depending on Stripe + eIDAS 2.0 mapping.

4. Competitive landscape

| Competitor | What they do | How we win | |------------|-------------|------------| | TruePic | C2PA-compliant signed photos, US-focused. | EU residency, NL-language reports, cheaper per-photo. | | ProofMode (Guardian Project) | Open-source signed photos. | Hosted, no setup, B2B-grade billing. | | Camify / TimeStampCamera | Visual timestamp overlay only. | Cryptographic signature + public viewer + tamper detection. | | Stock iPhone camera + manual SHA-256 | Free. | Workflow is 8 steps; we make it 1. | | In-house tools at large insurers | Some have internal apps. | Faster to adopt for SMB insurers and independent adjusters. |

The defensible moat is not the cryptography (a 200-line library) — it's:

The public verifier with a memorable URL, which adjusters and claimants link to from claim files.
eIDAS 2.0 conformance (qualified electronic signature mapping) once we publish that work.
AI-manipulation detection running on the server as a second-opinion layer.

5. Pricing

Three tiers; see issue #58 for the implementation breakdown.

| Tier | Monthly price (EUR) | Photos / month | Public viewer branding | Private photos | Retention | Bulk export | API | |------|---------------------|-----------------|------------------------|----------------|-----------|-------------|-----| | Free | €0 | 20 | "Powered by AuditPic" footer | ✗ | 30 days | ✗ | ✗ | | Pro | €12 | 500 | No footer | ✓ | 2 years | CSV / ZIP | ✗ | | Business | €49 | 5 000 | Custom domain for viewer | ✓ | 7 years | + API access | ✓ | | Enterprise | from €299 | custom | Custom domain + white-label | ✓ | configurable | + API + SSO | ✓ + SLA |

Annual plans get 2 months free (16.7% discount) — added later (issue #59 out-of-scope).

Why these numbers

Pro at €12/mo is in line with single-seat SaaS (Notion Plus €12, Zoom Pro €13). Adjusters expense it without procurement involvement.
500 photos covers ~25 inspections at 20 photos each — typical month for a freelance adjuster. Anyone over 500/mo is full-time and can justify Business.
Business at €49/mo undercuts TruePic's $99/mo while offering EU-resident storage.
Free at 20 photos is enough to evaluate but not enough to use full-time. Captures lead-gen value without giving the product away.

6. Unit economics (draft)

Per paid Pro user, monthly:

| Line | EUR | |------|------| | Revenue (net of VAT, NL B2C) | 9.92 | | Stripe fee (1.5% + €0.25) | (0.40) | | MinIO storage (avg 500 photos × 3 MB × €0.02/GB-month) | (0.03) | | Backend compute (Spring Boot on shared k3s) | (0.40) | | AI detection (one inference per photo, ~€0.001 per photo, 500/mo) | (0.50) | | Email / SMS notifications | (0.05) | | Customer support amortised | (1.50) | | Gross margin | 6.74 | | Gross margin % | 68% |

Customer Acquisition Cost target: ≤ 3 months payback → ≤ €20 blended CAC. LTV at 18 months retention → €120 — gives LTV:CAC > 6, which is healthy for SMB SaaS.

Sensitivity: AI-detection inference cost is the line most likely to move; if a usage-based GPU provider charges €0.005/photo we're at 60% gross margin instead of 68%. Worth re-modelling once volume is real.

7. Go-to-market

Months 1–3 — manual

5 design-partner accounts from the founder's NL insurance network. Free Pro for 6 months in exchange for a video testimonial and a 30-min interview about workflow.
LinkedIn outreach to NL/BE insurance loss adjusters (~1 500 prospects). 2-touch sequence + free-tier signup link.
Bouwend Nederland membership (construction sector body) — pay for a booth at one event.

Months 4–6 — content + partnerships

One detailed case-study blog post per design partner, indexed for terms like "fotobewijs verzekering," "schaderegeling foto authenticiteit."
Insurance-broker partner program: 20% recurring referral fee for 12 months. Brokers email their network.

Months 7–12 — paid + sales

Google Ads on long-tail queries ("expertise foto authenticatie," "fraude schadefoto").
One outbound SDR (€60k all-in) hitting tier-2 insurers + larger inspection firms.

Channels we explicitly are NOT using

Influencer marketing.
App-Store ASO as the primary channel (B2B audience doesn't search the App Store; they search Google).
Trade shows beyond the one Bouwend Nederland test.

8. 12-month revenue forecast

| Month | Free signups | Pro paid | Business paid | MRR (EUR) | |-------|--------------|----------|----------------|-----------| | 1 | 50 | 5 (design partners, €0) | 0 | 0 | | 3 | 200 | 25 (5 paid + 5 design) | 1 | 65 | | 6 | 700 | 90 | 5 | 1 325 | | 9 | 1 500 | 200 | 12 | 2 988 | | 12 | 3 000 | 380 | 28 | 5 932 |

Base case ARR exit 12 months: ~€71k. Pessimistic case 0.5× (if NL-only and slow content cycle). Optimistic case 1.6× (if one insurer pilot converts to an enterprise contract).

These are bottoms-up from 1 LinkedIn message → 0.5% reply → 25% trial → 6% paid conversion. Recheck monthly.

9. Key risks

AI-detection false positives. If our backend flags a legitimate photo as manipulated, we damage the very trust we're selling. Mitigation: tunable confidence threshold, human-review fallback for the Business tier, and a clear "this is advisory, not legal evidence" disclaimer until the model proves itself over 6 months.
eIDAS 2.0 implementation timeline. If the regulatory framework slips, our "qualified attestation" differentiation slips too. Mitigation: don't price on it; treat it as upside.
GDPR Article 30 (ROPA). Storing photos that contain identifiable people = PII. We need a clear processor agreement + data-residency story. NL-only storage on TransIP plus an SCC-compliant DPA solves this for EU customers. Tracked in ocs/auditPic/.
Camera-app spoofing on jailbroken devices. The system camera can be hooked to return library photos. Mitigated by issue #52 (jailbreak detection) and the in-flight #51 (camera-only).
Stripe / MinIO single-vendor risk. Stripe outages would block signups; MinIO failures would block uploads. Both have credible failover stories (Adyen / S3-compatible) but we don't implement until ARR justifies it.
Founder bandwidth. Until the first paid hire, sales + product + support is one person. Hard ceiling on month-12 revenue is the founder's calendar.

10. Milestones

| Quarter | Goal | Success metric | |---------|------|-----------------| | Q2 2026 | Free tier shipping on TestFlight + Play | 50 free signups; 5 design partners onboarded | | Q3 2026 | Stripe billing live; first paid customer | 10 paying Pro users; €120 MRR | | Q4 2026 | Public-viewer scale-up + case study #1 published | €1 500 MRR | | Q1 2027 | First Business-tier customer (broker / enterprise) | €5 000 MRR | | Q2 2027 | Hire #1 (customer success + content) | €10 000 MRR; CAC ≤ €25 | | Q4 2027 | NL → BE → DE expansion | €25 000 MRR; 2-country footprint |

11. What this plan deliberately doesn't decide yet

C2PA conformance — adopt as a marketing standard once the IPTC ecosystem coalesces. Don't engineer for it speculatively.
Per-photo pricing — risks confusion; revisit only if subscription churn correlates with under-usage.
Mobile-first vs web-first verifier — the public viewer is web-first today and stays that way; the mobile app is capture + own-photo verify only.
Open-sourcing the verifier — could be a trust signal; weigh against giving competitors the reference implementation. Decide after 100 paid customers.

Update this doc at the end of every quarter. Anything in here that turns out wrong is more valuable to revise than to retain.

Reacties

Nog geen reacties

Athena — auditPic/business/business-plan.md

AuditPic — Business Plan

1. What AuditPic is