Everybody is building the AI report. Nobody is building the proof.

The hardest thing in this market is already built — and nobody can find it

ScribeCam captures a field walkthrough — continuous video, photos taken mid-recording, spoken narration — and produces a finished report sealed on the device with SHA-256, a Secure Enclave signature and an RFC 3161 timestamp from an independent authority. Anyone can verify it for free, forever, without a ScribeCam account. No other product in the field-documentation market does this.

The foundation underneath it is empty. There is no App Store or Google Play listing, so there is currently no way for a buyer to install the product. There is no analytics, pixel or tag manager on any of the ten marketing pages, so no acquisition spend can be measured or optimised. There is no structured data of any kind, and robots.txt, sitemap.xml and llms.txt all return 404 — at the exact moment the company is trying to become an AI-cited entity. And 1,181 documented use cases, the single largest content asset in the business, are trapped on one URL.

[ Data verified ] Full crawl of scribecam.com and app.scribecam.com; ~90 live US keyword volumes via Search Atlas; YouTube data via vidIQ; LLM visibility across ChatGPT, Gemini, Perplexity, Copilot and Google AI Mode (Search Atlas project 5c1ecc68); Apple iTunes Search API and Google Play verification — all pulled 21 August 2026.

The seven questions, answered

A genuine category-of-one, described as a commodity

Verified from a full crawl of every public page — not from the pitch. ScribeCam, Inc. is a Delaware corporation headquartered in Santa Rosa, California (per its own Terms of Use; venue Sonoma County). The mechanic: continuous video recording while the user works, photos captured without stopping the video, spoken narration as the primary input, on-device transcription, sessions recorded per area and joined into one report. The output is two artefacts from one capture — a branded PDF and a gated online report on its own link — plus a ReportBundle.zip containing everything.

The seal: SHA-256 per file (FIPS 180-4), hardware-backed P-256 device signature from the Secure Enclave, itemised manifest.json, root fingerprint, RFC 3161 timestamp from an independent public TSA, and a permanent Bundle ID registration — written to the Federal Rules of Evidence 902(13)/(14) pattern. Verification is free, permanent and account-free, and works even if ScribeCam ceases to exist.

Four things competitors cannot copy

• Intellectual honesty in the evidence claim. The court-grade page separates captured evidence from derived material, states that AI writes the report but never the evidence, and says outright that no vendor can promise admissibility. That restraint is rare, and it is a genuine sales asset with lawyers and carriers.
• Zero-interruption capture. Every competing tool requires the user to stop observing and start entering. The argument that this destroys detail — not just time — is the sharpest thing on the site, and it is buried on /whyscribecam.
• The 1,181-use-case corpus. Seventy categories mapped to real job names with role lists per category. Most competitors have six vertical pages.
• Open standards, no lock-in. Ordinary zip, published standards, verification without the vendor — a procurement and legal-review accelerator almost nobody else can offer.

Messaging inconsistencies found across pages

The three-layer message stack

New categories fail when they lead with the abstraction. This stack leads with the pain, sells on the outcome, and closes on the moat. Every asset should be able to name which layer it operates on.

The demand is not in the category — it is in the competitors’ names

Roughly ninety live US keyword pulls across the entire concept space ScribeCam occupies. Add up every “field documentation,” “AI report,” “voice-to-report” and “proof-of-work” phrase and the whole category totals under 2,000 US searches a month. Several of the terms the product is literally built around return zero.

[ Data verified ] Search Atlas keyword database, US, pulled 21 August 2026. KD = keyword difficulty (0–100).

Pool 1 — incumbent brand demand (~340,000 US searches/mo)

The largest addressable pool by an order of magnitude. People are not searching for the problem; they are searching for the products they already know. Comparison and alternative pages intercept them at the exact moment they are evaluating.

Pool 2 — template and format intent (~4,000 US searches/mo)

The best-qualified pool in the entire plan. A person searching for a report template is, by definition, a person who currently writes reports by hand. That is the ideal customer describing their own pain in a search box.

The market, and the order to attack it

The qualifying test is mechanical: does a person physically travel somewhere, observe or alter a physical condition, photograph it, and then have to tell a third party what they found and what they did? ScribeCam has already documented more of this universe than any competitor — 1,181 use cases across 70 categories. Priority is set by four weighted dimensions: documentation intensity (30%), economic value (25%), reachability (25%) and product fit (20%).

[ Method ] Scores are Bonsai Marketing analyst assessments against the defined rubric, informed by the live keyword and AI-visibility research. A prioritisation instrument, not measured market data. Re-score after the first ten paying customers, when real CAC and cycle-length data replaces judgement.

The economic argument

ScribeCam is not sold as software. It is sold as the recovery of an unbilled second shift that every field business is already paying for and none of them have on a line item. The canonical worked example, run against ScribeCam’s own published pricing:

Two dozen rivals on the mechanic. None on the moat.

A live LLM visibility scan across ChatGPT, Gemini, Perplexity, Copilot and Google AI Mode surfaced 171 competing domains in ScribeCam’s answer space. A large share are AI-native field-reporting startups that already exist and are already being cited: ticnote.com, fieldreportai.com, videolens.io, fieldbinder.ai, struct.ai, energent.ai, complycam.com, inspectmind.ai, fieldscribe.report, anyinspect.ai, amnis.ai, fieldvox.com, buildsync.ai, spacecapture.ai, sitecam.com, brenix.ai, architai.com, voxreports.com, fieldnotesai.com, terrassist.ai, insytetracker.com, thincontext.ai and safeaspect.com.

And the incumbents are shipping the same mechanic as a feature. Fulcrum — the single most-cited domain in the scan at 12% of citations — already ships “Audio FastFill,” which lets field workers dictate observations hands-free and converts voice into structured field data. CompanyCam ships “AI Walkthrough Note,” which turns a walkthrough into a shareable report.

The gap ScribeCam sits in

Fifteen candidate positions, tested

Each was evaluated against three questions: is it defensible, is it unclaimed, and does it survive contact with a buyer who already owns CompanyCam?

Who to compete with, and who to partner with

Not every adjacent product is a competitor. ScribeCam produces a document; almost every platform below wants a better document attached to its job record and has no intention of building one.

A rare product on a foundation that does not exist yet

Ten launch blockers, and six of them are P0

These are not optimisations. Each one either wastes acquisition spend outright or damages credibility with the exact buyer ScribeCam is trying to win. Nothing in §07 onward should start until the P0 list is closed.

Site architecture — the highest-impact change available

Convert one page holding 1,181 use cases into a navigable, indexable, internally-linked hierarchy: fourteen /industries/ hubs, job-named /use-cases/ built in scored waves, seven /roles/, a /compare/ tree (Pool 1, build first), /integrations/, /roi/ calculators, /templates/ with real downloads, an /evidence/ section expanded from /courtgrade, plus /glossary/ and /research/.

Not a growth engine — an interception layer

Reframed by the first finding. SEO here is not demand capture: it is brand interception, template interception, and the long tail of 1,181 job names. It should be judged on assisted pipeline and AI-citation contribution, not on head-term rankings that lead nowhere.

Technical foundation to ship alongside

• robots.txt with sitemap declaration and explicit allow for GPTBot, ClaudeBot, PerplexityBot, Google-Extended and CCBot — ScribeCam wants to be trained on and retrieved.
• sitemap.xml, segmented by section, with accurate lastmod.
• llms.txt declaring the entity, the product, the proprietary nouns and the canonical explainer URLs.
• Self-referencing canonicals on every page.
• One title convention — recommend Primary Keyword | ScribeCam, retiring the mixed em-dash and trademark variants.
• Unified nav labels across both domains. “Use cases” beats “Who’s it for” on both search and buyer language.
• Analytics before any of it — GA4 with GTM, Meta pixel, LinkedIn insight tag, and conversion events for trial start, calculator completion, template download and demo request.

The long tail, generated rather than guessed

The validated corpus is ~90 terms pulled live. The full opportunity set comes from crossing three axes against the existing use-case library — job (1,181 documented cases, already written) × modifier (template, software, app, how to, checklist, example, requirements, form, report, best) × qualifier (for contractors, for insurance, free, court admissible, with photos, 2026).

[ Honest scoping ] This assessment reports the ~90 terms pulled with live volumes on 21 August 2026. The remaining expansion set is a defined method, not a list of numbers being asserted. The full validated set should be the first task after the P0 blockers close, because it also determines the build order.

Baseline zero — in a category where the answers are still up for grabs

Given the absence of search demand, this is not a supporting channel — it is arguably ScribeCam’s primary discovery channel. When a category has no search volume, buyers do not stop looking. They ask an assistant instead.

[ Data verified ] Search Atlas LLM Visibility, project 5c1ecc68, domain scribecam.com, US — twelve buying questions across four topics, measured across ChatGPT, Gemini, Perplexity, Copilot and Google AI Mode on 21 August 2026. Sentiment where mentioned: 75. Re-measure monthly; this is the section’s primary KPI.

Who owns the answers today

The five levers, in priority order

1. Entity resolution — no structured data exists anywhere on the site, so no machine knows what ScribeCam is. Ship Organization, SoftwareApplication, Product, FAQPage, HowTo, VideoObject, BreadcrumbList and Article schema, with sameAs pointing at every profile — Crunchbase, LinkedIn, both app stores, G2, Capterra, YouTube and the USPTO records. Then submit to Wikidata. This is the prerequisite for everything else.
2. Glossary — the highest citation-per-hour asset available. ScribeCam has a vocabulary nobody else is defining: ReportBundle, root fingerprint, tamper-evident capture, digital chain of custody, FRE 902(13), FRE 902(14), RFC 3161 trusted timestamp, Secure Enclave signing, SHA-256 manifest, derived vs captured evidence, sealed window, self-authenticating record. One clean 250–500 word definition page each, cross-linked.
3. Original research — four aggregator domains in this space exist purely to republish statistics, and assistants cite them constantly. Nobody owns the numbers in field documentation. Publish an annual State of Field Documentation report surveying 500+ field professionals. Every finding becomes a citable statistic with ScribeCam’s name on it, and it feeds PR, ads, outbound and sales decks simultaneously.
4. Third-party corroboration — models cite directories and roundups more than vendor sites. G2 and Capterra with 10+ reviews each, both app stores, Product Hunt, Crunchbase, the Zapier directory, and inclusion in the “best field documentation software” roundups that already exist and already omit ScribeCam.
5. Answer-shaped content — every target question gets an H2 that is the question verbatim, followed by a direct 40–60 word answer before any elaboration. That extractable opening paragraph is what gets quoted.

The ScribeCam AI Visibility Score

The channel missing from the plan — and it gates everything else

ScribeCam is a mobile-first product with no app store presence. For a tool a technician installs on a phone, the stores are not a distribution afterthought — they are the primary discovery surface and the entire install path.

Why this outranks most of the plan

• Every paid click, cold email and YouTube view currently terminates at a page with nothing to install. The conversion rate of the entire acquisition programme is capped at zero until this is fixed.
• App store search is intent-rich and free. “inspection report app,” “job photo app” and “field report” are typed directly into both stores by exactly ScribeCam’s buyer.
• Store reviews are a top-tier AI-citation source — apple.com already appears in ScribeCam’s citation set with 8 mentions.
• An enterprise buyer’s IT department will ask for the store listing before a pilot. No listing reads as no product.

Listing strategy

Zero reviews anywhere — a conversion problem, a diligence problem, and an AI problem

ScribeCam currently has no review presence on any platform. That costs installs, blocks enterprise diligence, and — per §08 — suppresses AI visibility, because review directories are among the sources assistants cite most heavily.

The case study machine

The single highest-value repeatable process available, because it gates the enterprise, PE, franchise and association motions simultaneously. Design it once and run it on every customer from customer one.

The one asset that unlocks enterprise

Commission a third-party evidentiary review: a recognised digital-forensics firm or a law-school evidence programme independently examines the ReportBundle and publishes an opinion on how it maps to FRE 902(13) and 902(14). ScribeCam’s own court-grade page is careful and honest about not overclaiming — which is exactly why an independent voice saying the same thing is so valuable. That single document will do more for enterprise, carrier and government pipeline than a year of advertising.

In a category nobody searches for, you do not capture demand — you create it

Google Ads — a precision instrument, not a scale channel

The brief asked for a full segmented paid-search programme. The live CPC data says something more useful: build it small, precise and competitor-led, because generic search in this category does not pay back at ScribeCam’s SMB price point.

Recommended structure, in build order: brand defence · CompanyCam competitor · Encircle/DocuSketch/magicplan · templates and formats · restoration intent · inspection intent · roofing storm intent (geo-triggered on NOAA hail and wind events, surging for 14 days) · remarketing · YouTube in-stream. Total test budget $13,600/month, run 90 days before judging. No Performance Max in year one — with no conversion history, no product feed and a tiny keyword universe it will spend the budget on broad-match junk. Exact and phrase only at launch. One landing page per campaign, never a shared homepage.

Meta — the primary demand-creation channel

Search finds people who already know they have the problem and are looking for a solution. §03 proved almost nobody in this market is doing that. Meta reaches people who have the problem, have never once thought of it as solvable, and recognise themselves instantly in three seconds of video. That is the definition of a demand-creation market, and it is where the budget belongs.

Budget: $8,000/month prospecting across three verticals plus $2,500/month retargeting. Hold lookalikes until 200+ trial signups exist — seeding off 40 conversions produces noise. Nine concepts at launch, three new per week thereafter: in demand creation the creative is the targeting, so assume 60–70% of variants fail and budget for volume. Every concept ships as an Instagram cut and a Facebook cut — Facebook is where the owners are; Instagram is where the twenty-eight-year-old lead tech who will actually champion the app is.

YouTube — promoted to a primary channel

Live YouTube data tells a completely different story from Google. The demand this category cannot find on Google is sitting on YouTube, at three to five times the volume, with materially lower competition.

[ Data verified ] vidIQ keyword research, pulled 21 August 2026, seed “inspection report software.” Competition index 0–100, lower is easier. Only the home-inspection cluster was pulled; expect the same pattern across restoration, roofing and adjusting — verify before committing budget per vertical.

Three conclusions. One: “home inspection software” is 3.6× bigger on YouTube than Google at a competition score of 10.6 — close to an open goal. Two: this is a watch-first buyer; field professionals evaluate tools by watching somebody use them, which is unusually good news for a product whose entire thesis is that it demonstrates itself. Three: YouTube is also the #15 citation domain in ScribeCam’s AI-search space, so every video is simultaneously a discovery asset, a sales asset and an AI-visibility asset. No other channel does all three.

Content architecture: Tier 1 vertical demos (8–12 minutes, one per Tier-A vertical, real job, real site, including the parts that need editing) · Tier 2 job-specific how-tos (3–6 minutes, mapped one-to-one onto the use-case pages and embedded there) · Tier 3 Shorts (three per week, cut from Tier 1 and 2 footage plus technician UGC) · Tier 4 the proof explainer (one genuinely excellent 10-minute film on how the seal works — low view count, extremely high value, and the most linkable and citable asset ScribeCam will ever publish) · Tier 5 “ScribeCam vs CompanyCam, honestly.”

Cold email — the highest-volume controllable pipeline source

In a category with no search demand, outbound is not a supplementary channel. It is how the market finds out the category exists.

Infrastructure before a single send: three to five separate sending domains — never scribecam.com — with full SPF, DKIM, DMARC and a custom tracking domain each, two to three mailboxes per domain, and 21–28 days of warm-up ramping to a hard ceiling of 30–40 sends per mailbox per day. Verify every address before it enters a sequence; bounce above 3% ends the programme. Route every positive reply into the CRM within minutes — outbound dies on slow response far more often than on bad copy.

Benchmarks to hold to: bounce under 3% · open 40–55% · positive reply 2–5% for a well-segmented list in these trades · meeting-booked 0.8–2%. Judge on positive reply rate, never open rate — open tracking is unreliable and rewards the wrong behaviour. Below 1% positive reply after two weeks means the list is wrong, not the copy. Volume ramp: 200/day in weeks 1–2, 500/day in weeks 3–4, 1,000–1,500/day from month two, only if bounce and reply metrics hold.

[ Planning figures ] List sizing by vertical — restoration ~12–18k firms, home inspection ~25–30k, field adjusting/IA ~4–6k firms, roofing ~40k+, property management ~15k firms, plumbing/electrical/HVAC ~120k+. These are Bonsai Marketing planning figures for list-build sizing, not verified counts.Needs confirmation Confirm against the chosen data provider before budgeting seats.

The AI SDR

Its real job is not sending more email — it is deciding, per company, which sequence and which message layer to use, at a quality no human SDR team could sustain across 100,000 prospects. Ten steps: discover → enrich → analyse their site → classify (vertical, tier, buyer role, message layer — the step that makes the whole thing worth building) → score → compose → orchestrate email and LinkedIn on one timeline → handle reply → CRM and retarget → learn.

LinkedIn outbound runs coordinated with email on a single timeline, never as an independent channel — the multi-touch effect only exists if the prospect experiences it as one company reaching out, not two. Day 1 profile view → day 2 email → day 4 connection request with no pitch → day 6 email → day 9 message → day 12 email → day 16 genuine engagement with their content → day 19 the breakup email → day 26 a final soft message. Hard account limits: 20–25 connection requests and 40–50 messages per sender per day. LinkedIn restrictions are near-impossible to appeal, and a restricted founder account costs more than the pipeline it was chasing.

Where one relationship equals hundreds of seats

Private equity and consolidators — the most powerful motion available

One relationship with an operating partner can put the product into forty branches at once, and PE-backed platforms buy standardisation more readily than any other buyer in the market, because standardisation is literally their thesis.

A roll-up acquires twelve plumbing companies. Each documents work differently — some with photos in a shared drive, some with paper, some with nothing. At exit, a diligence team will ask about operational consistency and liability exposure, and the answer will be twelve different answers. ScribeCam turns that into one answer, and does it without asking any of the twelve to change their workflow, because the workflow is “keep doing the walkthrough you already do.”

The pitch structure: open on the portfolio, never the product (“You’ve acquired eleven companies. How many different ways do they document a job?” — nobody has ever asked them this and every one of them knows the answer is bad) → quantify the labour across the portfolio, leading with a high six or seven-figure number and then showing the inputs so it survives scrutiny → introduce the liability line by asking what the current documentation would look like if a five-year-old job produced a claim, and let the silence do the work → propose a single-branch 90-day pilot with agreed metrics, never a portfolio rollout in the first meeting → price the portfolio, not the branch.

Franchise — start at the bottom, not the top

A franchisor’s brand is damaged by the worst franchisee, not the average one, and documentation is where that shows: one location sends a professional report, another sends four blurry photos and a text message. The franchisor has no mechanism to fix this because they do not employ the technicians. A tool that produces an identical, branded, professional document regardless of who is holding the phone is a brand-standards instrument disguised as a productivity app.

Priority systems to evaluate: restoration (Servpro, Paul Davis, ServiceMaster Restore, 911 Restoration, Rainbow Restoration, PuroClean) · pest control (Terminix, Orkin, Truly Nolen, Mosquito Joe) · home services (Mr. Rooter, Mr. Electric, Aire Serv, Roto-Rooter, Benjamin Franklin Plumbing) · inspection (Pillar To Post, HouseMaster, WIN Home Inspection, AmeriSpec) · property services (Real Property Management, Chem-Dry, ServiceMaster Clean).Needs confirmation

[ Scope note ] Systems listed are candidates for evaluation based on category fit. Confirm current ownership, existing vendor relationships and technology mandates before outreach — several sit inside multi-brand holding groups where one relationship reaches many brands.

Associations — a distribution channel hiding in plain sight

In trades organised around certification, the association is the market. The YouTube pull showed “internachi” and “nachi” each drawing over 5,000 monthly searches — members search for their association by name more than they search for software.

Then, in descending order of value: quarterly webinars per Tier-A association (the association does the promotion, and every recording becomes a YouTube asset) · vendor directory listings (low cost, permanent, and a §08 corroboration signal) · newsletter sponsorship, tested before committing annually · speaking slots for the founder on the deniability thesis, never on the product — programme committees reject vendor pitches and accept genuine industry arguments · member-benefit pricing, which gives the association something to announce.

Priority bodies: IICRC · RIA · InterNACHI · ASHI · PLRB · NAIIA · NAPIA · CLM · NRCA · NARPM · IREM · NAA · PHCC · NECA · ACCA · NPMA · AGC · ABC · ASSP · IFMA · BOMA · SAMS · NAMS · DRI. CLM is where claims meets legal — the single best audience for the Layer 3 message anywhere.

Trade shows — the booth demo is the product

ScribeCam has an unusual advantage: it is one of very few products that can be demonstrated completely and honestly in ninety seconds, using the booth itself as the subject. Hand the visitor the phone. Ask them to walk twenty feet of the exhibit hall and describe what they see out loud — the carpet seam, the scuffed panel, the cable run that should have been taped down. Stop. Tap create. Hand them the finished report ninety seconds later, sealed, with their voice behind it and their name on it. Then email it to them: the lead capture and the demo are the same action, and they leave holding a document they made.

Score every candidate show 1–10 on seven dimensions before committing budget — attendee fit, decision-maker density, cost to compete, competitive presence, demo suitability (weighted higher here than for most vendors, because loud halls hurt an audio-first product), partnership potential, and honestly-estimated pipeline. Above 45 attend; 35–45 walk it first without a booth; below 35 skip.

ABM — do not open this motion yet

Almost nothing is built — and almost everything is winnable fast

The first five dimensions score current state (higher = stronger today) and carry a coral 12-month target tick. The last three score headroom (higher = more opportunity waiting to be captured) and carry no tick — there, the bar is the size of the prize. “Local SEO Opportunity” from the standard scorecard is reframed as App Store & Marketplace Opportunity: local SEO does not apply to a national mobile SaaS, and the stores are ScribeCam’s equivalent discovery layer. Scores are directional analyst assessments built from the verified findings in §06–§12.

Clear the blockers, prove the motion, then compound

What to do first — ranked by impact

The first five moves

• Ship the app to the App Store and Google Play. Nothing downstream converts until this exists.
• Install GA4 with GTM, the Meta pixel and the LinkedIn insight tag, and define the four conversion events.
• Resolve the beta-versus-priced contradiction and the 5+ seat rate-card ambiguity, and delete the iOS 7 claim.
• Move “They can fake voice, photos and video. They can’t fake yours.” from /courtgrade to the homepage — and decide deliberately whether ScribeCam leads with proof or with speed.
• Approve the /industries/ architecture so the 1,181 use cases can start earning.

Open questions that need a decision before execution

• When does the app actually ship? Everything in Phase 1 is gated on it. If the answer is more than six weeks out, the plan should be re-cut with a pre-launch waitlist and design-partner motion in Phase 1, and acquisition pushed to Phase 2.
• Is the published pricing final? The FAQ says it is not, and the 5+ seat rate card is ambiguous. Both have to be resolved before any calculator, quote or ad can reference price.
• Will you lead with proof or with speed? This assessment recommends proof as the differentiator and speed as the hook. It is a real strategic choice with different site, ad and pricing implications, and it should be made deliberately rather than drifted into.
• What is the trademark posture on the name? fieldscribe.report, sitecam.com, complycam.com and Google Play’s ScribeWare all occupy adjacent name space in the same category.
• Is there appetite for the enterprise motion? SOC 2, the evidentiary review and a 12–24 month ABM cycle are real investments. If the plan is self-serve SMB only, §12 comes out and the budget moves to Meta and outbound.