Trueness · plan

Trueness.ai enterprise business and pilot plan

Trueness should give a consumer brand and its agencies a useful synthetic research workspace—Panel, Verdict, Challenge, Talk, Ask the research and Export—supported by a human calibration programme with clear limits. Everyday usefulness is the product experience. The Calibration Record explains which particular claims the human evidence supports.

25 October 2026, enterprise continuation. The working buyer is a billion-dollar consumer business with recurring brand/creative decisions and an executive sponsor. No demand, price, research validity or production implementation is established by this plan. All financial figures are USD. The written brand platform is distinct from the planning diagrams; no visual identity has been selected.

Current scope and reconciliation

The initial programme covers three audiences in two markets. Per audience-market cell, the first full wave has 160 developmental and 240 sealed quantitative completes; later waves have 80 developmental and 240 sealed. Foundation full waves are quarterly: 1,200 completes per market initially, then 960, rather than a fixed 1,000 split across unlimited audiences. Each cell has 12 initial qualitative interviews. Larger scope and semiannual-versus-quarterly cadence are costed explicitly.

Audits are priced at a proposed $15,000 per cell, $75,000 floor. One cell is an audience × market × language × agreed task/instrument; four reserved concepts do not establish validity for four unrelated task families. A seven-cell audit is $105,000 with $90,250 planning direct cost. The model's $110,000 average blends seven/eight-cell engagements and has $94,167 average direct cost.

The proposed Foundation catalogue price is $600,000 annually. At the mid reference direct cost of $440,000, that is 26.7% mature direct margin before staffing effects and overhead. The model's $750,000 average contract value yields 41.3%; do not interchange the two prices. Approximately $1.1m ACV is needed for 60% mature direct margin at that cost base. Larger markets/audiences need added fieldwork, research, review and operating capacity, not just a price multiplier.

The formal Codex review of Claude's documents identifies remaining inconsistent offers, unsupported market/map claims and wording that exceeds the research design. The Claude review records the earlier project-model gaps that this enterprise iteration addresses. Read review dates and open findings before treating the entire package as commercially reconciled.

Read the package

Question Document
What is the enterprise proposition and investor narrative? Business plan, distilled investor plan
Who competes and what does public evidence support? Market landscape, pricing, enterprise budget hypothesis
What should the brand stand for? Written brand platform, following the reference brand-page hierarchy
What carries forward from previous work? Implementation inheritance, construction lessons
How do waves and claims work? Enterprise calibration method, primary evidence register
What could it earn and how much cash does it need? Financial case, editable workbook, cadence comparison
What happens in the anchor pilot? Two-quarter pilot, six-month cash budget
How is it run with limited administrative work? 23 operating procedures, record templates
What does a customer do every day? Everyday use, six-action screen specification
What is the platform and how do assistants connect? Platform, models/API/MCP
What did the cross-review find? Codex on Claude, Claude on Codex

Enterprise financial scenarios

Case Y1 revenue Y1 EBITDA Y2 EBITDA Y3 revenue Y3 EBITDA Funding including 20%
Low $636,667 −$1,510,833 −$1,702,375 $6,123,542 −$2,403,156 $8,576,838
Mid $1,267,500 −$1,292,500 $1,171,766 $24,092,578 $6,692,926 $2,892,244
High $2,216,667 −$1,040,833 $10,318,917 $62,896,667 $33,801,100 $2,176,760

New programmes signed per year are 3/6/10 low, 5/14/28 mid and 8/24/50 high; ACVs are $500k/$750k/$1m. One conditional anchor signs at M1; other first-year wins follow six-to-nine-month sales cycles. Service begins two months after signing; annual churn occurs at renewal. Bookings, revenue and cash are separate. No automatic audit conversion or extra rebuild revenue is assumed.

The model includes $600k/$750k/$900k central first-year platform engineering, paid founder/commercial capacity, whole research leads hired ahead of service, external review, fieldwork, production AI, delivery engineering, client/agency management and included model-change reserves. It assumes net 60 collection. The initial fieldwork wave consumes 41% of the first-year fieldwork allowance, with 59% over the remaining three waves; this front-loading matters to funding. EBITDA is unchanged by moving cash timing, but cash need increases.

The low case is a stop/reprice scenario, not an investment recommendation. The mid case requires up to 23 research leads by year three, with about 7.8 FTE each in delivery engineering and client management, plus commercial/reviewer capacity. The high case reaches 40 leads. Administrative automation can reduce founder burden; it does not eliminate this human delivery organisation.

At the mid case, six months of sales delay increases funding to approximately $3.85m; net 90 collections require approximately $3.79m. Licence revenue is a separate, unverified partner-activity assumption, tested with a no-licence case. Supplier quotes, actual pipeline and signed channel terms must replace these assumptions before commitments scale.

The larger six-audience/four-market cadence comparison shows why scope cannot be added cheaply. At $1.5m, first-year direct margin is about 21% with semiannual full waves and negative with quarterly full waves. Achieving 60% on that first-year scope requires approximately $3.0m/$4.4m respectively under the cost assumptions. This is a pricing/capacity test, not a forecast of willingness to pay.

Anchor pilot and operating priorities

Pilot: one $600,000 annual Foundation programme plus two seven-cell $105,000 audits. Over six operating months, contracted first-year value is $810,000, recognised revenue $410,000, cash receipts $255,000 and cash expenditure $1,287,300. Funding with 20% contingency is $1,238,760. It includes prelaunch and the specified enterprise platform build; a smaller concierge option would need a separate scope/budget.

Two Foundation waves and the audits require 7,680 quantitative completes, including 6,240 sealed observations across distinct cell-waves, plus 128 interviews. These cannot be pooled into one audience sample. Statistical tolerances are decided per claim and powered before fieldwork; there is no universal accuracy score. A 150-person market pulse is a drift screen, not a new full-wave validation.

Read-only MCP reaches the anchor's authorised agency by M3, with brand-controlled grants, revocation, audit and isolation checks. Panel/Verdict/Challenge/Talk/research/export remain easy to use. Exploratory reactions are clearly generated; historical validation does not automatically make each new response or fictional character valid. Track repeated human use, Panel-to-action, citations, latency, follow-ups, decisions and actual human time separately from statistical performance.

The 23 SOPs include enterprise contracting, full-wave procurement, agency enablement, executive reviews, licensee oversight and model-version change. Existing-scope model-change remediation/revalidation is included in the annual programme. If a reserve is insufficient, fund the obligation or agree a remedy; never silently pass an untested replacement or turn an included obligation into an automatic upsell.

Evidence and completion status

The prior workbench informed workflow and failure-mode lessons; confidential source data and identifying lineage remain private. Rights/legal review precedes reuse or the first product build. General experience does not transfer proprietary research ownership.

The enterprise workbook passes all eight retained test families, including independent recalculation of Excel formulas, annual cohorts, accounting/cash identities, initial-wave cash, input validation and stress cases. These tests do not establish demand, statistical validity or implemented security. The specification's product tests remain work to implement.

The Codex-owned enterprise update and formal review are complete. Claude owns corrections to its reviewed documents and publication of the combined package. Next evidence: scoped buyer commitments, supplier quotes, rights-cleared inputs, a named methodologist/reviewer/custodian, funded engineering and a frozen first protocol.

All documents

brand

finance

operations

product

research

review

strategy