# 07 — Realtor Hub Launch Decision Sheet

## Purpose
Convert one week-one Realtor Hub review report into a safe next decision: proceed narrow, simplify, pause, or gather more evidence.

This sheet is a local review artifact only. It does not authorize recruiting, outreach, publishing, checkout, lead capture, ads, account creation, terms acceptance, CRM/MLS work, private-data use, or any external system action.

## Inputs allowed
- Fictional/sample weekly report from the local Week-One Review Wizard.
- Tim-supplied notes from explicitly approved manual reviews.
- Pseudonymized counts only: review completion, useful outputs, blocked risks, confusion, and return intent.

Do not include real client names, brokerage/MLS records, private CRM exports, lender/vendor private details, transaction files, account credentials, payment/tax/legal/KYC/identity details, or unsupported public claims.

## Decision lanes

### A — Proceed narrow
Use only if:
- 4/5 reviewed participants complete the weekly review.
- Useful outputs are specific enough to edit and use manually.
- Return intent is credible.
- Approval gates are understood and accepted.
- No unauthorized external action occurred.

Next local action:
- Draft participant criteria, data policy, and invitation copy for Tim review only.
- Keep scope to Today Board, Appointment Prep, Draft Desk, weekly review report, and approval wall.
- Defer team controls, integrations, public campaign assets, checkout, CRM/MLS sync, and automated sending.

### B — Simplify week two
Use if:
- Value appears, but onboarding is confusing.
- Participants like pieces of the tool but cannot explain the first-week routine.
- The review takes too long or too many modules compete for attention.

Next local action:
- Cut week two to Today Board + Appointment Prep + one reviewed follow-up draft.
- Rewrite onboarding so the first useful report appears in under 7 minutes.

### C — Pause safety
Use if:
- Demand centers on autopilot sending, scraping contacts, CRM/MLS writes, unsupported lead promises, public claims, or compliance shortcuts.
- Users resist the approval wall.
- Risky outputs would become client-facing without review.

Next local action:
- Do not launch.
- Turn risky requests into blocked-risk examples.
- Reposition around safer speed, templates, manual review, and approval checklists.

### D — Gather more evidence
Use if:
- Feedback is only “cool idea” or “looks useful.”
- Nobody completes a repeat review.
- No one says they would return or pay.
- The sample is too small or the notes are not specific.

Next local action:
- Collect 3–5 Tim-approved manual review notes before building more.
- Log exact objections, confusion, useful outputs, and return-intent language.

## Human approval lock
AI may draft, organize, score, summarize, route, remind, and create local review files. Tim/humans approve participants, data sources, facts, claims, messages, public copy, recommendations, testimonials, publishing, sending, lead capture, checkout, paid ads, account creation, terms acceptance, CRM/MLS/system changes, compliance-sensitive statements, and every external action.

## Forbidden from this artifact
- No recruiting or outreach.
- No public posting, publishing, sending, or scheduling.
- No lead capture, checkout, payment, paid ads, analytics, deployment, account creation, or terms acceptance.
- No scraping or private CRM/MLS/brokerage/lender/vendor/buyer/seller/transaction data.
- No CRM, MLS, email, SMS, calendar, ad platform, brokerage, lender, vendor, or transaction-system writes.
- No legal, tax, lending, pricing, disclosure, contract, agency, fair-housing, endorsement, testimonial, advertising, or compliance-sensitive claims without qualified human review.

## One-line answer
The safe next move is: proceed narrow / simplify week two / pause safety / gather more evidence.