Repeatable playbooks. Each SOP gets run multiple times. Lock the steps so the work doesn’t drift.
When: 3rd Wednesday of each month · 7pm Eastern
Pre-call (T-7 days): 1. Send theme email to cohort. One question, not a lesson. 2. Confirm Calendly RSVP count. Reach out personally to anyone who hasn’t responded. 3. Prep the prompt deck — 3 questions for breakouts · 1 reflective close.
Pre-call (T-1 day): 1. Tech check: camera · mic · lighting · recording setup if recording. 2. Re-read the cohort’s last journal entries (members who shared). 3. 30-min prayer before the call. No exceptions.
Call (90 min): - 0–10 min: Welcome · roll call · scripture opener (1 verse · KJV/NIV blend) - 10–25 min: Sandi opens the theme · 8-min teaching · 5-min reflection prompt - 25–55 min: Breakout (3 women per breakout · 30 min · 3 prompts) - 55–75 min: Re-gather · share 1 thing each (NO obligation) - 75–85 min: Sandi closes · prayer · next month preview - 85–90 min: Buffer for questions · scheduling
Post-call (T+1 day): 1. Send recording link (Tier 3 only). 2. One-question follow-up: “What landed?” 3. Update private notes: who showed up · who didn’t · who needs a check-in.
Post-call (T+7 days): 1. Review notes · flag anyone who hasn’t engaged in 30 days for a personal check-in.
When: Every Sunday evening
Steps: 1. Pull list of cohort members + their
assigned accountability partners. 2. For each pair: confirm they’ve
connected this week. If not, gentle nudge to one or both. 3. Update
accountability log (in
filled-instances/cohort-tracking/<cohort-N>.md). 4.
Flag patterns: women going dark for 2+ weeks → personal Sandi
outreach.
When: Automated · triggered by Stripe webhook on purchase
Steps: 1. Stripe webhook → IONOS PHP handler → MySQL
purchases record. 2. Cron job (every 6 hours) checks for
pending welcome emails. Sends Day 0 immediately on signup. 3. Cron
checks purchases.created_at against email queue: - Day 3
trigger - Day 7 trigger - Day 14 trigger - Day 30 trigger (Tier 1 →
upgrade · Tier 2 → next-step · Tier 3 → cohort confirm) 4. Each send
writes to email_log table with status (sent · bounced ·
opened if pixel).
Manual override: - If Sandi wants to pause sequence for someone (e.g., they emailed back asking to slow down) — admin panel toggle to suspend their queue.
See also:
_skills/email-sequence-builder/SKILL.md for content;
04_system_architecture.md for technical schema.
When: First week of each quarter
Steps: 1. Pull KDP analytics — last 90 days sales ·
search-term performance · sponsored-ads CTR. 2. Identify 3 winners and 3
losers from the 5 ad variants. 3. Run
_skills/kdp-ad-copy-writer/SKILL.md for 5 new variants on
the next test. 4. Update KDP description if the data shows the current
one underperforms. 5. Refresh A+ Content if season demands (e.g., Q4 →
gift positioning). 6. Don’t change the book title or subtitle. Don’t
change the cover. Stability matters.
When: Triggered when Sandi puts a book in beta phase
Steps: 1. Drop all feedback (emails · DMs ·
annotated PDFs · voice memos · in-person notes) into
projects/<book>/beta-feedback/. 2. Wait until at
least 3 readers have responded (ideally 5). 3. Run
beta-feedback-aggregator-agent (per AGENT.md). 4. Review
the consensus issues + theological flags. 5. Decide: structural rewrites
for consensus issues · defer single-reader concerns unless severity
demands action. 6. Pull praise quotes for KDP description + A+
Content.
See also:
_agents/beta-feedback-aggregator-agent/AGENT.md.
When: Before any draft ships (book chapter · email · blog post · ad copy · cohort prompt)
Steps: 1. Run draft through
_skills/voice-keeper-sandi/SKILL.md. 2. Address every BLOCK
violation (zero exceptions). 3. Address every WARN if the rewrite
improves the work. 4. NIT items optional. 5. Log violations to
_knowledge/VOICE_AUDIT_LOG.md (the calibration log).
No drafts ship without a voice gate. This is the cardinal rule of the brand.
When: Sandi wants to write a devotional that supports the KWB platform (e.g., a 7-day “Coming Home” mini-devotional offered as a free signup magnet)
Steps: 1. Pick week theme + anchor scripture. 2. Run
_skills/devotional-week-builder/SKILL.md for 7 days. 3.
Voice-gate. 4. Format for delivery channel (PDF lead magnet · email
autoresponder · in-platform locked content).
When: First Monday of January · April · July · October
Steps: 1. Pull metrics — book sales · Tier 2 conversions · Tier 3 active members · Tier 3 retention rate · cohort attendance avg · email open/click rates. 2. Flag any metric that dropped >15% QoQ. 3. Review tier price points. Adjust only if data demands. 4. Identify 1 thing to start in next quarter, 1 to stop, 1 to keep doing. 5. Schedule action items in PROJECT_STATUS_MASTER tier-promotion process.
When: Weekly (Sunday evening)
Steps: 1. Open 2026_Business_Receipts/
folder. Look at any files added since last sweep. 2. Use
receipt-watcher-agent to surface uncategorized receipts. 3.
For each: confirm category · vendor · amount · business %. 4. Append
rows to current month’s
filled-instances/receipts/2026-MM.md. 5. End of month: run
quarterly tax tracker update if quarter-end month.
When: Always
Rule: Sundays are NEVER work days. No platform deploys · no email blasts · no cohort calls · no KDP launches · no social posts auto-scheduled to fire on Sunday.
Implementation: - Cron jobs check day-of-week before any outbound action. - Drafts can be queued Saturday for Monday-morning send. - Auto-replies on email indicate Sandi will respond Monday.
This is non-negotiable. If a “great launch day” lands on Sunday, the launch waits 24 hours.
02_action_plan.md — when SOPs apply within the 90-day
window04_system_architecture.md — technical implementation
for SOP-03 (email automation)_skills/ — referenced skills per SOP_agents/ — referenced agents per SOP