DeckNarrate

1. One-liner

Dictate the boat survey out loud on deck; we hand back a defensible client-ready report before you drive home.

2. Trend signal — why now?

Independent marine surveyors say the report — not the inspection — is the job they hate. Direct industry sources:

• The International Institute of Marine Surveying (IIMS) describes the status quo bluntly: “The tedious task of taking handwritten notes and duplicating the effort by transcribing these notes into a legible document in the office is accepted as the unfortunate norm.” Surveyors report spending “the entire evening in front of the computer, even eating dinner at their desk,” with a nightly goal of transcribing “just half the notes collected from that day’s assignment.”
• Seaworthy Marine Survey states the report “can take anywhere from 4 to 12 hours or more” and that “the inspection is only about half the job.” A 35ft cruiser = 10–12 hours total.
• IIMS also flags the quality failure of the current generation of report software: reports with “minimal information given and the use of words that are inherently vague” — “adequate,” “apparently,” “ample,” single-word assessments like “poor” — boilerplate that “wouldn’t withstand legal scrutiny.”
• Existing software (InspectX at $42/report, Force5 at $50/mo, 3D Inspection) is template/data-entry tooling. None of it does voice-to-narrative AI. The IIMS deep-dive on report software does not mention AI, voice, or auto-narration at all.
• IIMS notes even outsourcing to a human transcriptionist fails: “the review and correction time was still too long to justify the expense.”

What changed in the last 12 months: speech models now transcribe noisy, domain-jargon, outdoor field audio reliably, and LLMs can turn rough spoken observations into structured, defensible survey prose cheaply. That specific combination did not work two years ago. Incumbents are still selling tap-the-template tablet apps.

Provenance:

• Signal 1 (demand): Surveyors hate transcribing notes nights/weekends; report = 4–12 hrs/job — https://www.iims.org.uk/introducing-inspectx-a-new-tool-for-the-old-school/ and https://seaworthysurvey.com/faqs/ — 2026-05-18
• Signal 2 (feasibility): Incumbent report software is template-only; IIMS report-software analysis never mentions AI/voice — https://www.iims.org.uk/the-opportunities-and-threats-presented-by-report-writing-software/ — 2026-05-18
• Signal 3 (economic): Paid incumbents actively selling (InspectX $42/report, Force5 $50/mo); surveys priced $500–$2,000+ so surveyors have wallet — https://inspectxpro.com/pricing and https://force5survey.com/ — 2026-05-18 Category: Tech-unlock

3. The opportunity

The incumbent isn’t a competitor — it’s the laptop at 9pm. Surveyors do the boat work they enjoy, then lose an evening (or a weekend) re-typing handwritten notes into a Word/template document. The two existing software categories each fail differently:

• Template tablet apps (InspectX, Force5): kill some re-typing by making you tap structured fields on a boat, in the sun, on a moving dock, with wet hands. They speed data entry but produce exactly the vague boilerplate IIMS criticizes — because tapping a dropdown that says “Fair” is faster than writing why.
• Human transcriptionists: produce real prose but cost too much and the correction loop is too slow.

DeckNarrate exploits the gap: the surveyor talks — naturally, the way they’d describe a defect to a colleague — and AI produces detailed, specific, defensible narrative prose, structured to the report sections, with the surveyor’s photos slotted in. 10× faster than typing, and better quality than dropdown boilerplate because spoken description is richer than a tapped field. The wedge is doing the one thing both incumbent categories can’t.

4. Target market

• Primary customer: Solo and 1–3 person independent marine survey firms doing pre-purchase / insurance / damage surveys on recreational and small commercial craft. SAMS/NAMS-accredited and non-accredited. English-speaking markets first: US, UK, Australia, Canada, NZ, Ireland.
• Why they buy: “I love the survey, I hate the write-up. I lose every evening and most Saturdays to it. Buyers chase me for the report because their contract clock is running, and I’m still on hull blisters at midnight.” Time is their only inventory — every hour on transcription is an hour not surveying a paying boat.
• Rough TAM reasoning: SAMS ~1,000 worldwide; NAMS adds several hundred; UK small-craft surveyors ~220; plus non-accredited operators and AU/NZ/CA. Realistic English-market serviceable base ≈ 4,000–8,000 active small-craft surveyors. At ~$1,200 ACV that’s a ~$5–9M revenue ceiling in the beachhead before adjacent expansion. Tight but real for a 1–2 person bootstrap; not a VC market — correct for this portfolio.
• Why now for them: Buyers increasingly expect 24–48hr turnaround (rush fees already exist), squeezing the surveyor’s evening even harder, exactly as field-grade speech AI becomes good enough to absorb the work.

5. Product sketch (MVP)

• Mobile voice capture on deck: surveyor speaks observations per section (“moving to the transom — gelcoat crazing port quarter, roughly 200mm, cosmetic, photo”), works offline, syncs later.
• Photo capture inline with a spoken tag, auto-attached to the matching report section.
• AI turns spoken notes into structured, full-sentence survey narrative in the surveyor’s section schema (hull, deck, rig, engine, electrical, safety, valuation, recommendations).
• Defect register auto-built: every spoken issue extracted into a categorized findings/recommendations table.
• Surveyor-branded output: their logo, disclaimers, SAMS/NAMS formatting, exported to PDF/Word for client + insurer.
• Review screen: side-by-side audio snippet ↔ generated paragraph, one-tap edit. The surveyor stays the author of record.
• Reusable boilerplate library (standard disclaimers, methodology) so only the boat-specific narrative is generated.

6. AI angle — what’s load-bearing

Remove the AI and this is a dictaphone plus a Word template — i.e., the status quo that fails. The AI is the entire product: (a) robust speech-to-text on noisy, accented, jargon-heavy outdoor audio; (b) structuring free-form spoken observation into the correct report section and a normalized defect register; (c) expanding terse spoken shorthand into specific, defensible prose without hallucinating findings the surveyor didn’t state. That last constraint — faithful expansion, zero invented defects — is the hard, valuable part and the quality bar IIMS is implicitly demanding.

7. Localization angle (if any)

N/A — this is a global play, English-language markets first. The pain (report transcription burden) and the report formats (SAMS/NAMS/IIMS conventions) are near-identical across US/UK/AU/NZ/CA. Localization is not the wedge; the voice-to-narrative AI is. Non-English markets (e.g. France, Spain — large recreational fleets) are a later expansion lever, not a launch requirement.

8. Business model — path to $1M–$5M ARR

• Pricing: $99/mo solo tier (unlimited reports) or $39/report pay-as-you-go for low-volume surveyors. Anchored above Force5’s $50/mo template tool — justified because we kill the evening, not just the re-typing.
• ACV: ~$1,200 (blended; many surveyors do 6–20 surveys/mo and will take the flat tier).
• Rough math to $1M ARR: ~850 surveyors × $99/mo × 12 ≈ $1.0M.
• Rough math to $5M ARR: ~3,500–4,000 paying surveyors (most of the English serviceable base) + a firm tier ($249/mo multi-surveyor with QA review queue) + per-seat expansion at small firms.
• Expansion path: firm/multi-seat tier; QA reviewer mode (senior surveyor approves junior dictation); adjacent verticals with identical “talk-the-inspection” shape (yacht delivery captains, equipment appraisers, commercial dive inspectors).

9. Go-to-market wedge — first 100 customers

• Scrape the directories. SAMS (marinesurvey.org), NAMS (namsglobal.org), IIMS, and marinesurveyor.com publish full member directories with names, regions, emails. That’s ~2,000+ addressable solo operators with public contact info. Cold email a 90-second screen recording: their typical handwritten note → finished narrative report. Expect 3–5% reply on a pain this acute.
• IIMS / SAMS channels. IIMS publishes the Report magazine and runs CPD/seminars; they already platformed InspectX with a feature article. A guest piece + conference demo lands directly in the buyer’s lap. Mike Wall’s “Report Writing for Marine Surveyors” seminar audience is a pre-qualified list.
• Boating-forum surveyor threads. Surveyors actively post on The Hull Truth, Sailboat Owners, Trawler Forum defending turnaround times. Engage where they already complain about evenings lost to reports.
• Referral loop. Surveyors are a small, gossipy guild within regional chapters. A free month for a referral that converts spreads fast in a 1,000-person society.

10. Build complexity — justification

Medium. Speech-to-text and LLM structuring are off-the-shelf APIs; the report schema, branded PDF/Word export, and offline-first field capture are standard mobile/web work. The genuine engineering is the faithful-expansion guardrail (specific prose, zero invented findings) and reliable transcription of noisy outdoor jargon audio — both need iteration with real surveyor recordings, not a research breakthrough. Pair ships a credible v1 in ~10–14 weeks.

11. Gating checklist

Gate	Pass?	Note
Legal in target market	✅	Productivity tool; surveyor remains author/signatory. No regulated activity performed by us.
Ethical — no harm / dark patterns	✅	Faithful-expansion constraint is core; surveyor reviews every paragraph against audio.
Market exists (evidence above)	✅	Paid incumbents, explicit verbatim pain, directories of buyers.
1–5 person team can build this	✅	Off-the-shelf AI + standard stack; 10–14 wk v1.
Launchable with <$50K / ₹40L	✅	Solo/pair build, API usage costs, no capex.

12. Feasibility score

Axis	Weight	Score	Notes
Problem intensity	20	16/20	Hair-on-fire for the surveyor’s evenings/weekends; documented verbatim across independent sources. Not quite 17+ because it’s a recurring tax, not a money-on-fire-today emergency.
Demand evidence	15	11/15	Strong: explicit complaints, paid incumbents, buyer turnaround pressure. Held below 13 — direct “I’d pay for a voice version” demand is inferred, not yet observed.
Build feasibility	15	11/15	Off-the-shelf AI; real work is the no-hallucinated-findings guardrail and noisy field audio.
Distribution clarity	15	11/15	Named directories with emails + a warm institutional channel (IIMS/SAMS). Conversion math still unproven.
Revenue mechanics	15	12/15	Pricing anchored to existing paid tools; ACV credible; clear flat-tier logic. Ceiling is modest but real.
Time to first revenue	10	7/10	Cold email to acute pain + pay-as-you-go option → paying users in 4–8 weeks of launch.
Defensibility	10	5/10	Execution + workflow lock-in (their schema, their boilerplate library) + niche brand. Copyable by an incumbent in ~12 months; data of corrected expansions compounds.
Total	100	73/100

13. Qualitative modifiers

Founder-fit tags

technical-heavy (speech + LLM guardrail engineering is the product) · domain-expertise-required (must learn SAMS/NAMS report conventions and what makes a finding legally defensible — a surveyor advisor is near-mandatory).

Key assumptions to validate (3–5)

1. Assumption: Surveyors will dictate naturally on deck and trust AI-expanded prose if they can review it against audio. How to test: 8–10 ride-alongs / recorded-note sessions; generate reports; ask “would you send this?”
2. Assumption: The faithful-expansion bar (zero invented findings, legally defensible specificity) is reachable with current models + guardrails. How to test: Blind-grade 30 generated reports with a senior surveyor against the IIMS vague-language checklist.
3. Assumption: ~$99/mo clears willingness-to-pay vs. Force5’s $50 template tool. How to test: Price-anchored interviews with 25 directory surveyors; pre-sell 10 annual seats.
4. Assumption: The English serviceable base is large enough (4,000+) to reach $1M+ comfortably. How to test: De-dupe SAMS/NAMS/IIMS/marinesurveyor.com directories into a real addressable count.

Risk flags

1. Market size: Narrow niche. Fine for a bootstrapped sub-$5M ARR business; fatal if you expected more. Adjacent “talk-the-inspection” verticals are the pressure-release valve.
2. Incumbent fast-follow: InspectX/Force5 already own the relationship and could bolt on voice. Mitigant: ship the quality+speed combo first, lock in via surveyor-specific schema/boilerplate and corrected-expansion data.
3. Liability/trust: A hallucinated defect in a legal document is catastrophic for the surveyor. The product must make over-claiming structurally hard and keep the surveyor unambiguously the author. This is both the top risk and the moat.
4. Platform dependency: Reliant on third-party speech/LLM APIs for cost and quality.

14. Structured verdict

Score:                  73/100
Verdict:                GO
Confidence:             Medium
Best-fit builder:       Technical founder (speech/LLM) + practicing marine surveyor advisor
Time to revenue:        4–8 weeks post-launch (pay-as-you-go + acute pain)
Capital to launch:      ₹4–8 lakh / $5–10K (API usage, directory outreach, no capex)
Top 3 assumptions to validate first:
  1. Surveyors trust reviewed AI-expanded prose — 8–10 recorded-note sessions, "would you send this?"
  2. Zero-invented-findings bar is reachable — blind senior-surveyor grading of 30 reports vs IIMS vague-language checklist
  3. $99/mo clears WTP vs Force5 $50 — price-anchored interviews + 10 pre-sold annual seats
Kill criteria:
  - Abandon if <3 of 10 ride-along surveyors say the generated report is send-ready after one review pass
  - Abandon if senior surveyor flags invented/over-stated findings in >10% of blind-graded reports and guardrails can't close it
  - Abandon if <2% reply rate across 500 directory cold emails to this acute a pain

15. Next step — 1-week validation sprint

• Day 1–2: Pull SAMS/NAMS/IIMS/marinesurveyor.com directories; de-dupe to a real addressable count and 30 hot prospects. Build a clickable demo from one real surveyor’s handwritten field notes → finished narrative report (no product yet).
• Day 3–4: Get 6–10 surveyors on calls. Show the demo. Ask: would you dictate instead of type? Would you trust it after review? Would you pay $99/mo? Record exact objections.
• Day 5: Decide. Go only if ≥5 of ~8 say “I’d switch” and ≥3 will pre-pay a deposit for early access. Otherwise no-go — the pain is real but the trust barrier or wallet isn’t, and a niche this size can’t absorb a weak conversion rate.

The falsifiable result: signed deposits from real, directory-sourced surveyors — not “that’s cool.”