Not everyone trusts AI yet. We built the plan for that.
Diffusion-of-innovation research is clear: only the first 16% of any market finds a new product on its own. The other 84% - the early majority, the late majority, the skeptical - are reached by a person they already trust, showing them something real. This is that plan: two engines, funded entirely from unit economics, with zero dollars spent on ads.
A retired dairy farmer watched a machine renew her truck registration, then signed up her whole church
Ellensburg, WA - county fair Saturday, 2028
She did not read a landing page or click an ad - there wasn't one. A local agent she already trusted, standing at a folding table between the 4-H barn and the pie contest, showed her the one thing that mattered: her registration, renewed, receipt in hand, in under a minute. She paid nothing to see it. Three weeks later she was still using it, and the woman who showed her got paid because of that - not because of the demo. This is the plan for the other sixty percent of American households: the ones a website was never going to reach.
One near-free, one self-funded - never a marketing budget.
The referral flywheel
In progressRogers' innovators and early adopters - roughly the first 16% of any market, the people who already trust AI and find products like ours on their own.
Why: This audience is nearly free to reach - they just need a reason to bring a friend. Dropbox grew 3,900% in 15 months with no ad spend at all on the strength of this loop alone.
How: A double-sided reward - real cash or Puppy-generated token credit - for both the referrer and the household referred. Paid only after the referred household is still active at day 30, never at the moment of signup.
Funded by: Product margin, the same way Dropbox funded storage and PayPal funded cash - a cost of goods, not a marketing budget line.
The 🤫 Community Agent network
In designRogers' early majority, late majority, and the skeptical - the roughly 84% of households a website, an ad, or a stranger's cold knock will not move.
Why: The Cooperative Extension Service reached every county in America with zero advertising by putting one trusted local human in each county, showing a working result instead of a pamphlet. The promotora model proves the same mechanic overcomes deep institutional mistrust. Trust transfers from the messenger, not the message.
How: We do not build this messenger network from nothing - we already own it. Certified Doorstep technicians demonstrate live, in the home or at a trusted community venue. Licensed 🤫 Yellow Pages professionals - the insurance agents, financial advisors, and healthcare workers already trusted by these exact households - become the referral layer, earning a receipted fee for an introduction their own client already trusts.
Funded by: A retention-gated appointment fee and commission, paid from first-year subscription margin - see the comp formula below. Never a marketing budget line.
Six layers, in the order a household experiences them.
The system that decides whether an agent may be dispatched into a given municipality at all - solicitor permits, background-check status, and Do-Not-Knock registry sync, checked before every visit, not after.
Candidate sourcing - Yellow Pages professionals first, general Community Agents second - run through the same house-call training module Doorstep already built, plus the 1099-classification test.
Territory assignment, appointment scheduling, and visit routing - reusing the Doorstep consent protocol verbatim: written scope before the knock, checkable agent ID at the door.
Every referral code and every agent visit is itself a consent-scoped grant, visible to the household and revocable - the same receipt-ledger primitive as the rest of the product, never a hidden pixel or a device fingerprint.
The payout formula - appointment fee plus retention-gated commission, structurally zero-weighted on recruiting other agents. Published so it is auditable, not just policy.
Every activated household tagged by channel and self-reported prior AI sentiment at signup - so the referral flywheel's easy wins can never mask whether the hard 84% is actually being reached.
Where this plan could actually fail.
The honest section. Each problem is genuinely open, and each has a falsifiable finish line - if any of these read as easy, we wrote it wrong.
The Koscot boundary, enforced by the formula - not by policy
A compensation system that pays an agent for recruiting other agents, rather than for real product sales to end users, is legally a pyramid scheme under the FTC's Koscot standard - regardless of intent. Policy promises don't survive an audit; the math has to.
Finish line: A payout formula whose partial derivative with respect to "agents recruited" is exactly zero, and whose only nonzero term is retained real households served - published as a literal formula on this page, not asserted in a handbook.
19,000 municipalities, no national sales-ops team
Door-to-door solicitation law is set city by city and county by county - permits, background checks, fees, curfew hours, and Do-Not-Knock registries all vary, across more than 19,000 incorporated US municipalities.
Finish line: A permit-status database keyed by municipality that gates dispatch - an agent cannot be scheduled into a territory without a green compliance flag - with the current coverage percentage published honestly, starting near zero.
1099 classification that survives an audit
IRS and DOL worker-classification tests turn on behavioral control, financial control, and the nature of the relationship. Structuring Community Agents as independent contractors while directing exactly how and when they work invites misclassification liability at scale.
Finish line: A documented classification test applied and passed before Phase 1 launch, re-validated on every change to the comp structure or field-operations playbook.
Attribution without surveillance
We need to know a referral or a Community Agent visit led to a retained household - without building a tracking system that violates the consent-first principle on the very channel meant to build trust.
Finish line: Attribution runs on the same receipt-ledger primitive as the rest of the product: a referral code or agent ID the household can see and revoke, never a hidden identifier.
The Sunrun failure mode, made structurally impossible
Door-to-door solar's own numbers show what happens when incentives reward closing over honesty - reporting on "dirty tactics" became a national reputational liability for the whole channel, not just the bad actors. One viral incident can undo years of trust in a news cycle.
Finish line: A published, hard-zero guardrail on substantiated misleading-claims incidents; any single substantiated incident suspends that agent within 24 hours - reported quarterly, the same discipline as the zero-data-incidents gate on Doorstep.
Rural and low-connectivity dispatch and payout
The households we most need to reach - older, rural, skeptical - are disproportionately likely to have inconsistent cell coverage and thinner banking access. A system that assumes an always-online smartphone and instant digital payout fails exactly the segment it exists to serve.
Finish line: Offline-capable visit logging that syncs when connectivity returns, and a payout path that works without a modern banking app - direct deposit with a physical-check fallback.
Measuring the hard segment honestly, not the easy one
Blended adoption metrics let an easy referral quarter quietly hide a Community Agent channel that is not reaching anyone new - the entire point of this program is the households a website could never move.
Finish line: Every activated household tagged by acquisition channel and a single honest self-reported prior-AI-sentiment question at signup, reported as a segmented KPI, never blended into one aggregate number.
Honest market-capture denominators, not fabricated ones
Some ICP segments have a clean public registry (licensed insurance agents, RIAs, healthcare providers); others do not (UHNW family offices, state-by-state auto dealers). A team under quota pressure will quietly estimate the hard ones and report them with the same confidence as the enumerated ones.
Finish line: Every published capture percentage states its denominator's source and confidence tier - Enumerated or Modeled estimate - and the two are never averaged together into one blended number.
Agent payout = f(retained real households only).
Payout(agent) = Σ over households h personally served by agent, of [ $25 · completed(h) + 0.15 · Y1Value(h) · ( 0.34·retained30(h) + 0.33·retained90(h) + 0.33·retained180(h) ) ]No term in this formula is a function of how many other agents this agent recruited, trained, or manages. Recruiting another agent pays exactly $0, always. This is the entire Koscot-safety mechanism - not a promise in a handbook, a property of the math.
Real formulas and design constraints, labeled as such.
| $25 / appointment | Community Agent fee for a completed, qualifying in-home or venue demonstration | benchmark - in line with Vector Marketing's $20-28 industry standard, funded from margin |
| 15% of Y1 value | Retention commission, paid in three installments gated on the household still active at day 30 / 90 / 180 | design constraint - never paid at the moment of sale |
| $0 | Any agent's earnings attributable to recruiting other agents | invariant - the Koscot-safety line, by construction of the formula |
| < $175 | Fully-loaded cost ceiling per retained household via the Community Agent channel | target - versus Sunrun's $36-39K per subscriber for a far higher-ticket product; ours must be proportionally tiny |
| ~19,000+ | Incorporated US municipalities, each a separate permit jurisdiction | physics of the rollout - why coverage is paced municipality by municipality, not claimed nationally on day one |
| 2.5% / 13.5% / 34% / 34% / 16% | Rogers' adoption-curve segments - innovators, early adopters, early majority, late majority, laggards | the referral flywheel reaches the first ~16%; the Community Agent network exists for the other ~84% |
| Day 30 / 90 / 180 | The three retention gates every payout - referral or Community Agent - must clear before a cent is paid | invariant - the single mechanism that keeps this program out of the door-to-door dirty-tactics trap |
What this plan will never do.
- No payment, ever, for recruiting another agent - built into the formula, not asserted as policy.
- No visit without a checkable agent ID and the published Doorstep consent protocol.
- No purchased contact or lead lists; no calls to numbers on the National Do Not Call Registry.
- No dispatch into a municipality without a green permit-compliance flag.
- No false urgency, no impersonating a utility or government worker, no visits to addresses on a local No-Knock registry.
- No blended channel reporting - referral and Community Agent metrics are always published separately.
One Launcher, Regional Sales Executives, one owner per ZIP.
Modeled on how Uber actually scaled with almost no HQ headcount growth: a traveling Launcher proves a region, then hands it to a locally-grown Regional Sales Executive, while exactly one ZIP Owner is accountable for each ZIP's sales and revenue.
The Launcher (one of the 4)
Owns: One new region at a time - never two regions launching simultaneously with a 4-person HQ.
Accountable for: Personally proving the playbook in a region's first 3-5 ZIPs before handing it off - the only role that ever touches a brand-new market directly.
Earns: Salary - the only true employees in this whole model are the 4 HQ people.
How they get there: Modeled on Uber's traveling city launcher: fly in, hire and train the first ZIP Owners, prove the numbers, then move to the next region once a local Regional Sales Executive can run it.
Regional Sales Executive
Owns: 50-150 ZIPs across one state or metro; oversees roughly 10-15 ZIP Owners.
Accountable for: The region's aggregate revenue and retention number - reviewed weekly by the Launcher during ramp, monthly once self-running.
Earns: 3% override on the Y1 subscription value of every retained household in the region, regardless of who sold it - real revenue only, never a recruitment bonus.
How they get there: Grown from within, almost always - promoted from the region's own best-performing ZIP Owner, the same way Uber's city GMs were frequently operators who had already proven themselves locally, not outside hires.
ZIP Owner
Owns: Exactly one ZIP code - the single accountable human for that ZIP's sales and revenue attainment.
Accountable for: Solo at first: personally runs the demos and closes. Once volume justifies it, recruits and trains their own Community Agents within that one ZIP.
Earns: The base formula directly on self-sold households ($25 + 15% Y1 value, retention-gated), plus a 5% override on what their own recruited agents sell - never a cent for recruiting the agent itself.
How they get there: Selected and trained by the Launcher during a region's first wave, exactly against the criteria below - nothing else.
Selection criteria, at every tier - deliberately never age, sex, or appearance
- • Lived or worked in the ZIP or region for 2+ years - local roots, not relocated for the role.
- • Passed the Doorstep + Household Reach certification with a top-quartile score.
- • A track record of retained households (for internal promotions) or an equivalent trust-building track record (community organizing, healthcare, education, financial services).
- • Never: age, sex, appearance, or any protected-class-correlated criterion - a firing-line violation of company policy, not a guideline, at every tier of this org.
The quota cascade
Monthly ZIP target (retained households) ÷ 4.33 weeks = weekly target ÷ agents working the ZIP = daily target per agentA ZIP with a 30-household monthly target and 3 people working it (the Owner plus 2 agents): 30 ÷ 4.33 ≈ 7 households/week for the ZIP ≈ 2.3/week per person.
Every quota is a retained-household number (day-30 gate already applied), never an appointment or knock count - so the cadence never rewards volume over the retention gate the compensation formula already enforces.
Stand-up: appointments run, demos completed, blockers - 10 minutes, same time, every day.
ZIP-by-ZIP review against the weekly target; the ZIP Owner owns the number, not excuses for it.
Retention-gated results close the books - a household counted this month is one that actually cleared day-30, not one merely signed.
The top 100 wealthiest US ZIP codes - sourced, not guessed
Method: Ranked by median or aggregate adjusted gross income (AGI) per ZIP code from the IRS Statistics of Income (SOI) public dataset, published annually and free to use - the same defensible, citable source a real ops team would build from, not a blog listicle.
Why: Puppy One's higher tiers and the burst-to-cloud audience (engineers, creators, and businesses with real token-generation needs) clear unit economics fastest in these ZIPs first - the same bottoms-up-by-wealth logic already used to sequence the UAE and insurance-directory rollouts.
The ranked list itself is a Phase 1 deliverable generated fresh from the current IRS SOI release - not asserted here, so the list is never stale or wrong on this page.
What % of the real market have we captured - and what's still open to hunt.
Market capture % (ZIP, segment) = Retained ÷ Addressable ICP
Capture(z,s) = RetainedAccounts(z,s) / AddressableICP(z,s) · OpenOpportunity(z,s) = AddressableICP(z,s) - RetainedAccounts(z,s) - ActiveInPipeline(z,s)Published per ZIP, per segment, never blended across segments of different confidence tiers. The Open Opportunity list is the literal, ranked hunting list handed to the ZIP Owner - the entities in the public registry that are not yet retained and not yet in an active pipeline.
| ICP segment | Denominator source | Confidence |
|---|---|---|
| General households in target ZIPs | Census ACS 5-year household counts and median income by ZCTA | Enumerated |
| Licensed insurance agents & producers | State DOI / NIPR producer registries - the same source already seeding the 🤫 Yellow Pages (638k+ producers, all 50 states + DC) | Enumerated |
| Registered investment advisers & wealth advisors | SEC IAPD (individual adviser reps) + FINRA BrokerCheck (broker-dealer reps) | Enumerated |
| Doctors' offices, hospital staff & healthcare practices | CMS NPI Registry (National Provider Identifier) - practice address by ZIP | Enumerated |
| Automotive sales (dealerships & salespeople) | State DMV dealer-licensing lists / NADA dealer directory | Modeled estimate |
| Ultra-high-net-worth family offices | No clean public registry exists - sized from RIA/wealth-advisor referral density in the ZIP, not enumerated directly | Modeled estimate |
Quarters, not vibes - county by county, the way it actually gets done.
Comp formula finalized and legally reviewed against the Koscot standard; permit-compliance data model designed; recruiting pipeline opened for the first Yellow Pages professional cohort in the pilot metro. Entirely in our control.
1099-classification test documented and passed; first 10-20 Yellow Pages professionals onboarded as referral partners in King County; the first solicitor permit obtained in the pilot municipality.
First free-venue demo events - a library, a senior center, a credit union - run by certified Doorstep technicians; first 100 Community-Agent-assisted households; day-30/90/180 retention reporting live and segmented by channel.
Permit compliance expands municipality by municipality within Washington state, the same way Cooperative Extension went county by county starting in 1906; coverage percentage published, not claimed.
Multi-state expansion sequenced strictly by permit-compliance readiness, never by ambition. "Every American household" is the press release we are working backwards from - and we say plainly it is not there yet.
The north star and its guardrails.
Retained skeptical households served
Retained households who self-reported skeptical-or-neutral AI sentiment at signup - the actually-hard segment this program exists to reach, not the easy referral wins.
Day-30 / 90 / 180 retention by channel
Reported separately for referral and Community Agent - never blended - so an easy quarter on one channel can never hide a stall on the other.
Municipalities with green compliance status
The honest measure of how much of the country we are legally allowed to reach today, published against the ~19,000-municipality total.
Active Yellow Pages referral partners
Licensed professionals actively introducing the product to their own clients, with a receipted fee on each retained referral.
$0 recruitment-weighted earnings
Published quarterly as a hard zero - the Koscot-safety invariant, proven, not promised.
Substantiated misleading-claims incidents
Target zero; any single substantiated incident suspends that agent within 24 hours.
Fully-loaded cost per retained household
Published against the < $175 ceiling for the Community Agent channel - the number that proves this is self-funded, not a hidden ad budget.
One is a product of Hushh Technologies Corporation (brand: 🤫 “hussh”), an independent company. One runs on third-party silicon, systems, and cloud; all company names are used solely to describe the platforms on which One software runs. Hushh Technologies is not affiliated with, endorsed by, sponsored by, or partnered with any company named.
Reached by a person you already trust.
If you are one of the licensed professionals in the 🤫 Yellow Pages, or you want to bring 🤫 to your own community, this is the plan we are building with you.
One is a product of Hushh Technologies Corporation (brand: 🤫 “hussh”), an independent company. One runs on third-party silicon, systems, and cloud; all company names are used solely to describe the platforms on which One software runs. Hushh Technologies is not affiliated with, endorsed by, sponsored by, or partnered with any company named.