STRATEGIC PLAN · v1.2

Become the operator who quietly owns 8 businesses by 2030 — because his cost structure is half his competitors'.

Deploy production AI coworker agents (Maple, Mickey, Goldie) across FlexStay, Donuts, Property, and Goldenglaze. Capture the labor savings. Use the savings as an underwriting edge to acquire SMB #5 by October. Document everything publicly to compound deal flow. Six months. Three tracks. One thesis.

Sprint Header

Operator: Jack Yen
Start: 2026-05-20
End: 2026-11-20
Businesses today: 4
Target by Nov 20: 5 (+#6 in LOI)
Annualized EBITDA target: $850k → $1.35M
Hours reclaimed/wk: 120 → ~25

Executive Summary

THE THESIS

Jack is not a SaaS founder. He's a multi-business operator with working AI infrastructure (Maple, Mickey, Goldie running on Discord + MCP). Highest-EV use of that infrastructure is internal margin capture + AI-leveraged acquisitions — not selling AI to other people.

THE EDGE

Working production agents that can absorb 20-30% of an ops-heavy SMB's labor cost within 14 days post-close. Sellers can't price this in. Brokers can't underwrite it. Jack can pay 3-3.5x stated SDE and be at 2.0-2.5x post-AI.

THE RISK

Single biggest failure mode: trying to also build the SaaS product. Don't. Six months of focus is the entire bet. Secondary risk: sourcing slippage in June kills the October close. Tertiary: AI savings thesis under-delivers (<15%).

Three Parallel Tracks

AProductionize agents

25 workflows across 4 businesses — Goldenglaze first, FlexStay second, Donuts third, Property fourth. ~95 hrs/wk reclaimed at autonomous saturation. Builds the asset that makes Track B underwriting honest.

BAcquire SMB #5

$1.5M-$3M revenue, $150-500k SDE, labor >25%. Pay 3.0-3.5x stated, achieve ≤2.5x post-AI. Close Oct 15. Acquisition #6 in LOI by Nov 15.

CContent engine

Document the work as it happens. Sunday Rollup as signature artifact. Audience compounds into deal flow + optional consulting. Byproduct, not job.

6-Month Timeline

Track	Jun	Jul	Aug	Sep	Oct	Nov
A · Ops	Baseline + FlexStay V-tier	Donuts / Property deploy	Charter promotions	Newton onboarding agent build	Day-1 install at SMB #5	Standardize "operator OS"
B · Acquire	SBA pre-qual + 20 NDAs	12 CIMs · 4 mgmt calls	First LOI sent	QoE + financing	Close (target Oct 15)	30-day proof + LOI #6
C · Content	Cadence locked (10/wk X, 3/wk LI)	First podcast booked	First case study published	3-5 podcasts	Acquisition diary peaks	5k-15k X · 1.5k-2k newsletter

Critical Path

QBO billing reconciliation (F3) unlocks Donuts vendor ingest + owner reporting.
Guest message triage (F1) unlocks damage workflow + classifier reuse.
Vendor dispatch primitives (F2) port directly into Property maintenance.
3 production agents are the input to Newton (the onboarding agent).
Newton v1 is the deliverable that makes Track B safe at scale.

Stop-the-Sprint Conditions

Any business below 60% automation coverage at Month 4 (Sept 20).
Agent error rate >3% of actions for 2 consecutive weeks.
Personal liquidity below $100K for >30 days.
SBA lender appetite tightens beyond P+3.5%.
Content time exceeds 6 hrs/week — flywheel has reversed.

PLANNER

Day-by-Day Action Plan

Use this as your daily driver. Open the current week, do the tasks for today, check them off mentally. June is fully day-resolution. Months 2-6 are weekly milestones anchored to specific dated events. Time blocks are suggested defaults — flex them, but keep the count of items per day.

P0 · Daily Operating Rhythm (Mon-Fri default)

Time	Block	What you do
6:30-7:00	Review	Discord #ops-metrics red/yellow/green skim · overnight escalations · today's Hostfully board
7:00-7:45	Track A	Resolve overnight agent escalations · approve queued V-tier drafts
7:45-8:00	Track C	Ship the 8am X post (scheduled previous day in Tweet Hunter)
8:00-12:00	Deep work	Track A build/deploy work (one workflow at a time) or Track B sourcing block
12:00-12:15	Track C	Ship the 12pm X post · 10-min reply window on X
12:15-1:00	Lunch	Off — actually leave the desk
1:00-5:00	Operating	Customer calls · vendor coord · whatever the 4 businesses need · Track B calls if scheduled
5:00-5:30	Inbound	Triage #inbound-triage Discord (DMs, emails, podcast invites)
5:30-5:45	Schedule	Queue tomorrow's 8am + 12pm X posts in Tweet Hunter · update Notion task tracker
Wed 6:00am	Cash flash	Wed-only: mid-week QBO pull (cash/A/R/A/P) auto-posts; review for surprises
Tue/Thu/Sat 8am	LinkedIn	Ship LI post (3/wk cadence) — same content, repurposed format
Fri 2-3pm	Reply day	30 min of replies to operator-builder content on X

P1 · Weekend Template (recurring every week)

Day	Time	What you do
Sat	9:00-10:00	Assemble Sunday Rollup draft (RollupBot drops it Fri 5pm; you edit, sanitize, fact-check)
Sat	10:00-10:30	Schedule Rollup for Sun 9am in Tweet Hunter · cross-post draft to LinkedIn
Sat	rest	Off
Sun	9:00	Sunday Rollup auto-ships
Sun	9:00-10:30	Weekly Ops Review (§A6): agent_action_log → QBO P&L → charter cards → 7 metrics verified
Sun	10:30-11:00	Plan next week — fill in the daily planner items, queue Monday 8am X post
Sun	rest	Off

P2 · June — Day-by-Day (Days 1-30)

Recurring items from the daily rhythm above are implicit. The list below is the specific work for each day on top of the rhythm.

Week 1 · Jun 1-7 · DEPLOY GOLDIE + LIGHT THE FUSE

Mon Jun 1 — Deploy Day

9-11a · Track A: spin up Goldie Discord bot identity (token, perms, channel ACLs); add to Goldenglaze Discord server
11a-12p · Track A: connect mcp__gmail-goldenglaze + mcp__quickbooks (Class=Goldenglaze); test pulls
1-3p · Track B: book SBA pre-qual call with Preferred Lender #1 (try Live Oak or Newtek); send PFS draft for review
3-4p · Track B: open BizBuySell Pro; create 3 saved searches (STR portfolio, small bookkeeping firm, multi-loc cleaning) with target-box filters
4-5p · Track A: scaffold agent_action_log Notion DB (B1) — schema per §A3, no n8n sync yet
X morning post: "Day 1. Six months. One thesis." (Pillar #4)

Tue Jun 2 — Wire Day

9-11a · Track A: G1 (support triage) — create gmail-goldenglaze label filters; draft Goldie classifier prompt v1; Notion FAQ DB seeded
11a-12p · Track A: G3 (social/review monitor) — Google Business Profile webhook test; manual scrape fallback if needed
1-3p · Track B: email 5 local biz brokers in your region (intro + 2-paragraph target-box; ask for off-market deal flow)
3-5p · Track C: stand up Hook Library Notion DB (B16); seed 15 starter hooks from §C3 examples
LI post: "How I'm running 4 businesses with an AI team" (Pillar #1)

Wed Jun 3 — Triage Day

6a · Wed cash flash auto-posts to #ops-metrics; eyeball for surprises
9-12 · Track A: G1 first Shadow-tier run; review 10 drafts; record approve-as-drafted%; tune prompt
1-3p · Track B: Apollo enrichment — pull 50 owner-operator prospects matching target box
3-5p · Track A: build n8n agent_action_log sync workflow (B1); test write path
X posts: morning Build Log + lunch Agent-did-this

Thu Jun 4 — Source Day

9-12 · Track A: G4 (inventory low-stock) — define SKU thresholds; alert logic; Discord channel for low-stock pings
1-2p · Track B: hand-write 5 cold letters to off-market owners (paper, signed, mailed today)
2-4p · Track B: review BizBuySell saved-search results (~15-20 listings); screen against §B4 red flags; tag winners in Notion DB
4-5p · Track A: tune Goldie persona prompt with first 3 days of learnings
LI post: Acquisition Diary entry — "Week 1 of search. Here's my target box." (Pillar #3)

Fri Jun 5 — Reply Day

9-11a · Track A: G5 (weekly P&L snapshot) — cron set for Sun 8am; output template; Discord post format
11a-12p · Track C: draft RollupBot template doc (no n8n yet — just the Notion template Jack will manually run for week 1)
1-2p · Track B: follow up with 3 brokers from Tue (ask for first listing match)
2-3p · Track C: Reply day — 30 min in replies to operator-builder posts on X
3-5p · Track A: write trust-threshold guardrail middleware spec (B8) — pre-send hook design
X post: Ops Teardown — first FlexStay savings number (anonymized)

Sat Jun 6 · Sun Jun 7 — Weekend Template (see P1)

Sat: assemble and schedule the first-ever Sunday Rollup (manual, 8-section template, real numbers even if small)
Sun 9am: Rollup ships · 90-min ops review · plan Week 2

Week 1 exit criteria: Goldie in Goldenglaze Discord ✓ · G1/G3/G4/G5 in Deploy or Shadow tier ✓ · agent_action_log DB live ✓ · 3 broker relationships opened ✓ · 50 prospects in Notion ✓ · 10 X posts + 1 LI post + Sunday Rollup ✓ · SBA pre-qual call scheduled ✓

Week 2 · Jun 8-14 · SCHEMA + SOURCING

Mon Jun 8

9-12 · Track A: B2 Postgres analytics mirror — provision DB, n8n hourly sync from Notion → Postgres
1-3p · Track B: SBA pre-qual call with Lender #1 (have PFS, last 3yr tax returns, FlexStay K-1s ready)
3-5p · Track B: open Track B Deal Pipeline Notion DB (B11) with full schema from §B8
X morning: Build Log — "How I structure my SMB acquisition Notion"

Tue Jun 9

9-12 · Track A: G3 (social monitor) — first 3 days of data in; calibrate sentiment classifier
1-3p · Track B: email 5 more local brokers (different region) + 5 industry-specific brokers (VRM Intel network)
3-5p · Track A: G1 promote — if approve-as-drafted% ≥80%, move from Shadow → Supervised tier
LI post: Ops Teardown

Wed Jun 10

6a · Wed cash flash
9-12 · Track A: start Scout agent (B13) — BizBuySell scraper, scoring against target box, post to Discord
1-3p · Track B: review NDAs received so far; sign 3-5, request CIMs
3-5p · Track C: Content Performance Notion DB (B14) — wire Twitter Analytics → Notion via n8n
X posts × 2

Thu Jun 11

9-12 · Track A: G4 inventory alerts — Shopify/commerce MCP wiring (build B6); test with 10 SKUs
1-3p · Track B: read 2 CIMs · log in Notion · score against §B4
3-5p · Track A: Goldie weekly review — what did it draft, what got escalated, what to tune
LI post: Acquisition Diary — "What kills a deal in the first 60 seconds of a CIM"

Fri Jun 12 — Reply Day

9-11a · Track A: B4 charter card automation — Notion live-compute promotion/demotion gates from Postgres
11a-12p · Track A: Trust-threshold middleware (B8) v1 — banned-phrase + dollar-amount pre-send inspection
1-2p · Track B: follow up with 2 brokers from Mon-Tue; ask "got anything off-market?"
2-3p · Reply day
3-5p · Track A: portainer-compose SHA bump if any flexstay-ops or n8n images changed this week

Sat-Sun · Weekend template

Sun: first metrics review — Goldie approve-as-drafted%, hours saved, agent_action_log count. First data point on the savings curve.

Week 3 · Jun 15-21 · NDA WAVE + FLEXSTAY EXPANSION KICKOFF

Mon Jun 15

9-12 · Track A: F6 (FlexStay calendar/pricing hygiene) — promote from S → V if shadow ran clean
1-3p · Track B: scout-driven CIM review batch (Scout has been running 4 days)
3-5p · Track A: G2 (fulfillment exceptions) — start build, deploy by Jul 1
X morning: Build Log

Tue Jun 16

9-12 · Track A: F1 promote from V → A if metrics qualify (§A4)
1-3p · Track B: 2 owner-direct cold-outreach replies → if any reply, schedule mgmt call
3-5p · Track C: first Substack draft (Pillar #4 essay — "Why I run AI in Discord not Slack")
LI post: Ops Teardown

Wed Jun 17

6a cash flash
9-12 · Track A: G5 weekly P&L — review after 2 weeks of auto-posts; numerical error rate
1-3p · Track B: target → 8 NDAs signed by end of week
3-5p · Track C: schedule Substack essay for Sun publish

Thu Jun 18

9-12 · Track A: F2 (cleaner/vendor dispatch) — start deploy phase for July
1-2p · Track B: customer reference call practice — script for when CIMs become live deals
2-5p · Track B: 3 more CIM reviews
LI post: Acquisition Diary

Fri Jun 19

9-11a · Track A: Goldie autonomy charter — first promotion attempt (G5 P&L snapshot to A-tier)
1-2p · Track B: weekly broker check-in cadence — text 5 brokers "any new listings this week?"
2-3p · Reply day
3-5p · Track A: portainer SHA bump check

Sat-Sun · Weekend template + Substack publishes Sun 7am

First Substack essay goes live Sun Jun 21 at 7am. Cross-post excerpt to LI Mon morning.

Week 4 · Jun 22-28 · MONTH-1 PROOF

Mon Jun 22

9-12 · Track A: F3 (billing recon) Deploy phase
1-3p · Track B: 2nd SBA lender call — get parallel pre-qual
3-5p · Track A: 30-day Goldie metrics rollup — hrs saved, error rate, calibration
X: "Goldie at 30 days: what worked, what didn't" (Pillar #1)

Tue Jun 23

9-12 · Track A: F1 fully autonomous decision — confirm or hold based on metrics
1-3p · Track B: review week's CIM batch · advance 1-2 to deep-dive (mgmt call)
3-5p · Track C: build first Operator Q&A piece — reach out to 3 operator-builders Jack respects
LI post

Wed Jun 24

6a cash flash
9-12 · Track A: AP/AR variance review across all 4 businesses (first cross-business consolidation)
1-3p · Track B: target acquisition criteria refinement — what's working, what's noise
3-5p · Track A: Newton design doc — start spec'ing the AI onboarding agent (B9 build is Sept, design now)

Thu Jun 25

9-12 · Track A: G2 fulfillment exceptions — first Shadow runs
1-3p · Track B: first deep-dive mgmt call (if a target qualified)
3-5p · Track C: draft RollupBot n8n workflow (B15) — pulls QBO + Notion + Discord
LI post: Acquisition Diary — "What I asked on my first mgmt call"

Fri Jun 26

9-11a · Track A: month-1 hours-saved calculation across all 4 businesses (proxy: action count × avg minutes)
11a-12p · Track A: charter card automation deploy (B4)
1-2p · Track B: weekly broker check-in
2-3p · Reply day · also pitch podcast #1 (Acquiring Minds) this afternoon
3-5p · Track A: end-of-month portainer SHA bump

Sat-Sun · Weekend + month-1 retrospective

Sun: month-1 retro Sunday Rollup — headline number, "what I learned in 30 days"

Month-1 exit criteria (Jun 30): Goldie at V-tier on 3+ workflows · F1 graduated to A-tier · 20 NDAs signed · 5-8 active deep-dive targets · agent_action_log + Postgres + RollupBot live · 50 X posts shipped · 1,000+ X followers · 80+ newsletter subs · SBA pre-qual letter in hand from 2 lenders · 1 podcast pitched.

P3 · Jul-Nov — Weekly Milestones

Daily rhythm continues unchanged. The list below is the specific weekly milestone and the anchor day for each — dates are exact; the supporting daily work is implicit from June's pattern.

July (Days 31-61) · FUNNEL FILL + ROLLUPBOT LIVE

Week 5 · Jun 29-Jul 5

Mon Jun 29: Begin FlexStay greenfield expansion — F2/F3 in Deploy
Tue Jun 30: Goldenglaze 30-day savings memo published in Notion (internal)
Wed Jul 1: RollupBot n8n cron goes live (Fri 5pm draft)
Thu Jul 2: CIM review target = 3 this week
Fri Jul 3: Pitch podcast #2 (Owned and Operated)

Week 6 · Jul 6-12

Mon Jul 6: Inbound DM Triage agent (B17) build start
Tue Jul 7: F3 billing recon → Shadow tier
Wed Jul 8: First management call (if a target qualified from Week 5)
Fri Jul 10: RollupBot autonomously drafts the Sunday Rollup for the first time — review carefully Sat

Week 7 · Jul 13-19

Mon Jul 13: First IOI sent on a deal that cleared deep-dive
Tue Jul 14: G1 customer support triage promotes to V-tier (target by mid-Jul)
Thu Jul 16: Substack essay #2 — Pillar #2 Ops Teardown with 60-day savings numbers
Fri Jul 17: Podcast #1 (Acquiring Minds) recording (if booked Jun 26)

Week 8 · Jul 20-26

Mon Jul 20: Donuts D1 (vendor invoice ingest) Deploy phase
Wed Jul 22: 2nd IOI sent
Thu Jul 23: Inbound DM Triage agent fully live
Fri Jul 24: Pitch podcasts #3 (Sweaty Startup) + #4 (SMB Attorney)

Week 9 · Jul 27-Aug 2

Mon Jul 27: Month-2 retro — must have 12 CIMs reviewed, 4 mgmt calls done
Wed Jul 29: First LOI drafted (target: send by Aug 15)
Fri Jul 31: Decision rule check: if <15 CIMs reviewed → sourcing is broken, pivot to direct outreach

Month-2 exit criteria (Jul 31): 12+ CIMs reviewed · 4+ mgmt calls · 2 IOIs sent · RollupBot drafting autonomously · podcasts #1 booked, #2-4 pitched · F1+F6 at A-tier · G1+G3+G5 at V-tier or higher · 1,600 X / 2,200 LI / 200 newsletter.

August (Days 62-92) · FIRST LOI + CASE STUDY

Week 10 · Aug 3-9

Mon Aug 3: Donuts D2 (daily sales close) Deploy
Tue Aug 4: First case study post draft — anonymized 60-day numbers (publish target: Aug 8)
Wed Aug 5: 3 deep-dive targets in active review
Fri Aug 7: Podcast #2 recording
Sun Aug 9: Case study cross-posted everywhere — Sunday Rollup, LI, Substack

Week 11 · Aug 10-16

Mon Aug 10: Customer reference calls on top deep-dive target (3-5 customers)
Thu Aug 13: Newton spec finalized — implementation begins Sep 1
Fri Aug 14: Engage QoE provider on best target ($8-15k)
Fri Aug 15: ★ FIRST LOI SENT

Week 12 · Aug 17-23

Mon Aug 17: LOI response window — if accepted, kick off full QoE
Tue Aug 18: G2 fulfillment exceptions promote to V-tier
Thu Aug 20: Substack #3 — Pillar #2 Ops Teardown with 90-day numbers
Fri Aug 21: Pitch podcasts #5 (MFM) + #6 (Get Paid For Your Pad)

Week 13 · Aug 24-30

Mon Aug 24: If LOI accepted: full diligence sprint — financial + legal + AI-readiness audit
Wed Aug 26: If LOI not accepted: send 2nd LOI on backup target
Thu Aug 27: First Operator Q&A post (Pillar #5)
Fri Aug 28: Hook Library refresh routine — retire underperformers, clone winners

Month-3 exit criteria (Aug 31): First LOI sent (Aug 15) · LOI accepted by month-end OR 2nd in flight · Case study published · QoE engaged · 3 podcasts recorded · F2/F3 at V · G2 at V · Donuts D1/D2 at Shadow · 2,800 X / 2,800 LI / 400 newsletter / 4 inbound/mo.

September (Days 93-122) · DILIGENCE + NEWTON BUILD

Week 14 · Aug 31-Sep 6

Mon Aug 31: Newton build kickoff (B9 — 40hr spec); F1 design transfer to Newton's pattern lib
Tue Sep 1: Acquisition equity reserve in segregated HYSA — must be $200K+ by today
Wed Sep 2: Property P1 + P3 Deploy phase
Fri Sep 4: Decision rule check: no LOI sent yet? → quality bar too narrow, recalibrate

Week 15 · Sep 7-13

Mon Sep 7: SBA underwriting begins (45-60 day clock)
Wed Sep 9: Newton week-2 milestone — interview pipeline working
Thu Sep 10: Substack mid-month — "The AI-leveraged underwriting model"
Fri Sep 11: Podcast #3 recording

Week 16 · Sep 14-20

Mon Sep 14: QoE report due back; review with attorney
Wed Sep 16: Donuts D3 (supplies reorder) Deploy
Thu Sep 17: Customer reference calls (5 customers, 30 min each)
Fri Sep 18: Day-1 integration playbook drafted (what gets done Mon after close)
Sun Sep 20: ★ Decision rule check: any business below 60% automation coverage = pause, don't acquire

Week 17 · Sep 21-27

Mon Sep 21: Legal close prep — purchase agreement redlines
Tue Sep 22: Newton v1 ready for first test interview (use an FlexStay or Goldenglaze re-onboarding as the test case)
Thu Sep 24: Pitch podcasts #7 + #8
Fri Sep 25: Funding finalization checkpoint with SBA lender

Month-4 exit criteria (Sep 30): Acquisition deal in late-stage diligence · funding 80% confirmed · day-1 playbook ready · Newton v1 functional · 5 podcasts cumulative · 4,500 X / 3,500 LI / 750 newsletter / 8 inbound/mo · all 4 businesses at ≥60% automation coverage.

October (Days 123-153) · CLOSE WEEK + DAY-1 INSTALL

Week 18 · Sep 28-Oct 4

Mon Sep 28: Final legal review
Wed Sep 30: SBA conditional commitment letter
Thu Oct 1: Buyer-seller walkthrough · key employee stay-bonus conversations
Fri Oct 2: Pre-close checklist 95% complete

Week 19 · Oct 5-11 — CLOSE WEEK PREP

Mon Oct 5: Final QC on close docs
Tue Oct 6: Wire transfer prep · escrow funding
Wed Oct 7: Day-1 install dry run with Newton (in sandbox)
Thu Oct 8: Existing-business automation verified at ≥70% — green light to acquire
Fri Oct 9: Final close meeting w/ legal + lender

Week 20 · Oct 12-18 — ★ CLOSE

Mon Oct 12: Signatures · wire · entity transfer
Tue Oct 13: ★★ Close target Oct 15 ★★ · Newton interviews new owner Day 0
Wed Oct 14: Day-1: MCP credentials connected (QBO Class added, Gmail, whatever PMS/CRM)
Thu Oct 15: Day-2: Maple/Mickey/Goldie-equivalent agent deployed in Shadow tier
Fri Oct 16: Day-3: first 24 hours of shadow data reviewed; X post — "Signed for SMB #5. Day 3."

Week 21 · Oct 19-25

Mon Oct 19: SMB #5 week-1 metrics baseline locked
Wed Oct 21: Promote first workflows from Shadow → Supervised
Thu Oct 22: Substack — "How we got AI-leveraged ops live in 14 days at SMB #5"
Fri Oct 23: First paid consulting inquiry triaged (only if pipeline ≥5 qualified DMs)

Week 22 · Oct 26-Nov 1

Mon Oct 26: SMB #5 first ops cost data point
Wed Oct 28: Acquisition #6 sourcing kicks off in parallel
Thu Oct 29: Podcast #5-7 recordings this month
Fri Oct 30: Month-5 retro

Month-5 exit criteria (Oct 31): SMB #5 closed by Oct 15 · ≥3 workflows in Shadow/Supervised on SMB #5 · day-14 case study published · acquisition #6 sourcing live · 7,500 X / 4,200 LI / 1,200 newsletter / 13 inbound/mo · 7 podcasts cumulative.

November (Days 154-184) · 30-DAY PROOF + THE 6-MONTH ROLLUP

Week 23 · Nov 2-8

Mon Nov 2: SMB #5 day-15 metrics: target -15% labor cost vs baseline
Wed Nov 4: All-portfolio dashboard view — 5 businesses, consolidated
Thu Nov 5: Newton v1 stress test — onboard "hypothetical SMB #6" in <4 hours
Fri Nov 6: First acquisition #6 NDAs signed

Week 24 · Nov 9-15

Wed Nov 11: SMB #5 day-30: -25% labor cost target — if missed, root-cause why
Thu Nov 12: Acquisition #6 first CIM under review
Fri Nov 13: Decision: keep going on #6 LOI or pause until SMB #5 fully proven
Sat Nov 14: Begin drafting the 6-month rollup Substack essay

Week 25 · Nov 16-22

Mon Nov 16: Acquisition #6 LOI sent (if signal is real)
Tue Nov 17: ★★ The 6-Month Rollup Substack essay ships ★★ — definitive long-form
Wed Nov 18: Cross-post excerpts across all channels for 5 days straight
Thu Nov 19: Podcast #11 + #12 recordings booked or recorded
Fri Nov 20: ★★ END OF SPRINT ★★ · final retrospective starts

Week 26 · Nov 23-29 — AFTERMATH

Mon Nov 23: Sprint debrief in Notion · what compounded, what to drop
Tue Nov 24: 2027 cadence decision: stay on X or invest YouTube long-form
Wed Nov 25: Year-2 plan kickoff document (separate doc, not part of this sprint)
Thu Nov 26-Fri Nov 27: Thanksgiving — actually take it off

Month-6 exit criteria (Nov 20 — END): 5 businesses owned · ~$300-500K annualized ops savings captured · SMB #5 EBITDA expansion thesis proven · acquisition #6 in LOI or diligence · ≥12 workflows at A-tier · Newton v1 production-ready · 10K-15K X / 5K LI / 1.5-2K newsletter / 6-8 podcasts cumulative / 15-20 qualified inbound per month · 6-month rollup essay published.

P4 · Stop-the-Sprint Rules (check every Sunday)

Jun 30: Goldie not at V-tier on any workflow → diagnose root cause, do NOT start FlexStay greenfield expansion
Jul 31: <15 CIMs reviewed → sourcing funnel is broken, escalate to direct owner outreach immediately
Aug 31: No LOI sent → quality bar too narrow, recalibrate target box (do NOT lower discipline)
Sep 20: Any business <60% automation coverage → pause Track B, fix Track A
Oct 15: Close slipping → ok to slip to Nov-Dec for quality; not ok to slip on diligence rigor
Nov 11: SMB #5 day-30 labor savings <10% → core thesis wrong, pause acquisition #6
Any time: personal liquidity <$100K for >30 days → forced-sale risk, pause Track B
Any time: agent error >3% for 2 weeks → trust collapse, pause new workflow deploys
Any time: Sunday Rollup missed → that's the cadence canary, don't skip a second

TRACK A

Productionize Agents Across All 4 Businesses

20 workflows, four agent personas, three tiers of autonomy. By Nov 20, ≥12 workflows at autonomous tier and ~95 hrs/week of human time reclaimed. Output: a transferable playbook that makes SMB #5 deploy in week 1, not month 1.

A1 · Rollout Calendar

Legend: D Deploy · S Shadow · V Supervised · A Autonomous

Priority order: Goldenglaze first (lightest agent footprint today, fastest to mature) → FlexStay second (production agents continue running; new workflows added) → Donuts third → Property fourth. Newton (cross-business onboarding agent) builds in Sep once the first two businesses are at V-tier.

#	Business	Workflow	Jun	Jul	Aug	Sep	Oct	Nov
★ Priority 1 — Goldenglaze (Goldie) · greenfield deployment
1	Goldenglaze	G1 Customer support triage	D	S	V	V	A	A
2	Goldenglaze	G3 Social / review monitor	D	S	V	A	A	A
3	Goldenglaze	G5 Weekly P&L snapshot	D	S	A	A	A	A
4	Goldenglaze	G4 Inventory low-stock alerts	D	S	V	V	A	A
5	Goldenglaze	G2 Fulfillment exceptions	—	D	S	V	V	A
6	Goldenglaze	G6 Reorder vendor comms	—	D	S	V	V	V
★ Priority 2 — FlexStay (Maple) · in-production agents continue; new workflows phased in
7	FlexStay	F1 Guest message triage (already V)	V	V	A	A	A	A
8	FlexStay	F6 Calendar hygiene + pricing flags	S	V	A	A	A	A
9	FlexStay	F2 Cleaner / vendor dispatch	—	D	S	V	V	A
10	FlexStay	F3 Billing reconciliation (QBO)	—	D	S	V	V	A
11	FlexStay	F4 Owner monthly reporting	—	—	D	S	V	V
12	FlexStay	F7 Review response drafts	—	—	D	S	V	V
13	FlexStay	F5 Damage / incident workflow	—	—	—	D	S	V
Priority 3 — Donuts (Mickey)
14	Donuts	D1 Vendor invoice ingest → QBO	—	—	D	S	V	V
15	Donuts	D2 Daily sales close + variance	—	—	D	S	V	A
16	Donuts	D3 Supplies inventory reorder	—	—	—	D	S	V
17	Donuts	D5 Reviews + DM triage	—	—	—	D	S	V
18	Donuts	D4 Staff schedule draft	—	—	—	—	D	S
19	Donuts	D6 Cash drop / petty-cash log	—	—	—	D	S	V
Priority 4 — Property (Maple-LT lane)
20	Property	P1 Tenant comms triage	—	—	—	D	S	V
21	Property	P3 Rent follow-up sequence	—	—	—	D	S	V
22	Property	P2 Maintenance dispatch	—	—	—	—	D	S
23	Property	P4 Lease renewal calendar	—	—	—	—	D	S
24	Property	P6 Move-in/move-out checklist	—	—	—	—	D	S
25	Property	P5 Owner reporting (if applicable)	—	—	—	—	—	D
Cross-business
26	Cross	Newton — AI Onboarding Agent v1	—	—	—	D	S	V

A2 · Workflow Detail (with exact metrics per workflow)

P1Goldenglaze — Goldie 6 workflows · ~12 hrs/wk reclaimed at A-tier · greenfield, lead deployment

Workflow	Trigger	Tools	Key Metrics	Hrs/Wk Saved
G1 Support triage	gmail-goldenglaze + DMs	gmail-goldenglaze, Notion FAQ, Quo	Response <4h · FAQ-deflection >50% · CSAT ≥4.5/5	~4
G2 Fulfillment exceptions	Carrier webhooks	Commerce MCP (build), gmail	Resolved <24h >90% · Replacement cost <2% rev	~3
G3 Social/review monitor	Hourly scrape + GBP webhook	Scraper, Notion, Discord	Mention-to-alert <30min · Neg-review response <12h	~2
G4 Inventory low-stock	Daily 7am	Commerce MCP, Notion, Discord	Stockout days/SKU <2/mo · Alert FP <15%	~1
G5 Weekly P&L snapshot	Sun 8am cron	QBO Class=Goldenglaze	Delivered 52/52 weeks · Numerical errors 0	~1
G6 Reorder vendor comms	From G4	Gmail, QBO PO	Vendor SLA tracked · Lead-time variance trend	~1

Why Goldenglaze first: smallest agent footprint today + lowest-stakes customer-facing decisions (refund <$40 threshold, no owner-money flows, no regulated comms). Pattern of "Shopify-style commerce + gmail + Quo + QBO" is also the most portable template for acquisition #5 if Track B lands a DTC target. Goldie maturing fast = Newton has more training data by Sep.

P2FlexStay — Maple 7 workflows · ~47 hrs/wk reclaimed at A-tier · already in production

Workflow	Trigger	Tools	Key Metrics	Hrs/Wk Saved
F1 Guest msg triage	Hostfully webhook (real-time)	Hostfully MCP, Notion msg DB, Discord #guest-drafts	Draft latency <90s · Approved-as-drafted 70% (mo3) → 85% (mo6) · Escalation <5% · Review score +0.15★	~14
F2 Vendor dispatch	Checkout + recurring	flexstay-ops MCP, Quo SMS	Lead time <4h · Unit-ready-by-checkin >98% · Reassignment <8%	~10
F3 Billing recon (QBO)	Daily 6am cron	QBO MCP, Drive archive	Statement→QBO match >95% · Month-close ≤4 days · Owner-payable variance <0.5%	~8
F4 Owner monthly reports	1st of month 8am	QBO P&L by Class, Gmail-property	Lead time <2 biz days · Numerical error 0 · Owner Q-response <48h >90%	~6
F5 Damage / incident	F1 classifier flag	Hostfully, flexstay-ops dispatch_issue_as_job, QBO	Time-to-evidence <24h · Deposit-claim win >70% · Missed insurance = 0	~3
F6 Calendar / pricing flags	Hourly cron	Hostfully list_*	Orphan-night precision >90% · Gap alerts <2h · Pricing FP rate <15%	~4
F7 Review responses	Hostfully review webhook	Hostfully, brand-voice Notion	Coverage 100% · Median <24h · Sentiment match >4/5	~2

P3Donuts — Mickey 6 workflows · ~18 hrs/wk reclaimed at A-tier

Workflow	Trigger	Tools	Key Metrics	Hrs/Wk Saved
D1 Vendor invoice ingest	Gmail-donuts label	gmail-donuts, QBO MCP, Drive	OCR accuracy ≥98% · Ingest <2h · Duplicate <0.5%	~5
D2 Daily sales close + variance	11pm cron	POS CSV (Drive), QBO, Discord	Close-by-midnight >95% · Cash-vs-deposit variance <$10/day · Forecast MAPE <12% by mo6	~4
D3 Supplies reorder	Mon/Thu 6am	POS depletion, gmail-donuts, QBO PO	Stockouts <1/mo · Emergency reorders <2/mo · Waste -15%	~3
D4 Staff schedule draft	Wed 4pm	Forecast model, Notion, Quo	Draft acceptance >70% · Understaffed shifts <1/wk	~3
D5 Reviews + DM triage	Webhooks / scrape	Manual then scraped, Discord	Coverage >95% · Response <6h · Recovery offers redeemed >40%	~2
D6 Cash drop log	EOD + 10pm cron	QBO journal, Notion	Logged within 24h 100% · Recon gaps <$5/wk	~1

P4Property — Maple-LT lane 6 workflows · ~18 hrs/wk reclaimed at A-tier

Workflow	Trigger	Tools	Key Metrics	Hrs/Wk Saved
P1 Tenant comms triage	gmail-property + Quo	gmail-property MCP, Notion tenant DB	First-response <2h · Misclassified urgency <5% · CSAT ≥4.4/5	~5
P2 Maintenance dispatch	P1 classifier / portal	flexstay-ops vendor pool (shared), Quo	Ack <1h · Dispatch <4h · Close <72h · Reopen <8%	~4
P3 Rent follow-up	9am cron, days 2/5/10/15	QBO invoices, Quo, gmail	On-time-rent +5pp · Days-late mean trend down · Template misuse = 0	~3
P4 Lease renewal calendar	Daily cron T-90/-60/-30	Notion lease DB, GCal, gmail	90-day notice misses = 0 · Renewal close-by-T30 >75%	~2
P5 Owner reporting	Monthly cron	QBO, gmail	Same shape as F4	~2
P6 Move-in/move-out checklist	Lease boundaries T-7	Notion, vendor MCP	Completion 100% · Deposit-return SLA compliance 100%	~2

A3 · Agent Action Log — Schema

Every agent action, every business. One row = one decision moment. Notion is the cockpit; Postgres mirror handles analytics JOINs.

Field	Type	Why this field
id	UUID	Idempotency for n8n retries; correlates to agent stdout
timestamp	ISO8601	Latency + cadence; partitions weekly rollup
business_id	Select (flexstay/donuts/property/goldenglaze)	P&L attribution; required for dashboards
agent_id	Select (Maple/Mickey/Goldie/Newton)	Per-agent reliability profile
workflow_id	Select (F1-F7, D1-D6, P1-P6, G1-G6)	Joins to autonomy charter
action_type	Select (draft/send/dispatch/post/recon/flag/escalate)	Distinguishes "thought" from "wrote to the world"
trigger_source	Text	Root-cause analysis when workflow misfires
tools_used	Multi-select (MCP names)	Tool-failure attribution
status	Select (drafted/approved/edited/rejected/sent/escalated/failed)	Core graduation metric (edited ≠ approved)
human_reviewer	Person	Reviewer-load balancing
time_to_human_decision_s	Number	Detects review-queue starvation
downstream_outcome	Select (resolved/reopened/escalated/refunded/cancelled)	7-day-lagged join: the only honest "did it work"
dollar_value_proxy_usd	Number	Converts hours to $; reservation value, invoice size, refund amount
confidence_score	0-1	Calibration check vs observed outcomes
model_version	Text	Regression detection on prompt/model changes
notes	Rich text	Jack's annotations during weekly review
case_study_flag	Checkbox	Tags rows worth keeping for content + investor decks

A4 · Autonomy Charter — Tier Gates

Tier 0 → 1

Deploy → Shadow

Zero connection errors over 7 consecutive days
All required MCP tools respond <2s p95

Tier 1 → 2

Shadow → Supervised

≥80% "would have approved" rate over rolling 14 days
Zero "would have caused harm" events
Classifier confidence calibration within ±0.1 of observed outcomes

Tier 2 → 3

Supervised → Autonomous

≥85% approved-as-drafted (not edited) over rolling 30 days
≤2% escalation rate
Zero financial errors >$50 and zero legal/PII incidents
Median time-to-human-decision <8 min
Downstream resolved rate ≥90%

Auto-Demotion

Autonomous → Supervised/Shadow

3 escalations in 7 days → demote to V
Any financial/legal/PII incident → demote to S immediately
Approved-as-drafted <75% over 30 days → demote to V
Single customer/owner complaint citing agent → manual review + likely demotion

A5 · Trust Thresholds — Human-Only Forever

These workflows never graduate to Autonomous regardless of metrics. The guardrail middleware enforces these pre-send.

FlexStay

Refunds >$50 · Owner-money decisions · Expenses >$200 to owner
Deposit claims (any size) · Damage liability statements
Insurance correspondence · Listing price changes >5%
Cancellation/relocation w/ credits · Any legal threat reply

Donuts

Vendor bills >$500 · New vendors · Cash writeoffs >$25
Payroll/schedule changes >4h/wk per employee
OSHA/health-dept · Refunds >$25 · Loyalty credits >$10

Property

Any eviction/notice language · Rent abatement (any size)
Security deposit deductions · Lease term changes
Maintenance >$300 · Habitability complaints · Insurance claims

Goldenglaze + Cross-Cutting

Refunds >$40 or 2nd refund same customer · Recall/safety
Influencer/PR replies · Vendor POs >$500
All: any msg mentioning lawyer, lawsuit, illness, injury, child, police, immigration, or media → auto-escalate to Jack

A6 · Sunday 9am Weekly Ops Review (90 min, recurring)

Dashboards (in order)

Notion agent_action_log weekly view (business × workflow × status)
Discord #ops-metrics Monday-7am roll-up (preview Sun)
QBO consolidated P&L — 4 classes side-by-side vs last week
Hostfully today-board via mcp__flexstay-ops__get_today_board
Charter cards Notion view — workflows within 5pp of promotion/demotion

7 Metrics to Verify Every Sunday

Approved-as-drafted % per workflow
Escalation rate per workflow (trend over 3 wks)
Hours saved (cumulative trajectory → 95 hrs/wk)
$$ value moved through agents (risk exposure)
Charter promotions/demotions this week
Trust-threshold breaches (must = 0)
Discord/gateway uptime + failover events

Questions Jack asks each agent in Discord (logged): Maple: "Top 3 escalations this week and what would have prevented them?" · Mickey: "Variance >8% days — root cause each?" · Goldie: "Refund requests — pattern?" · All: "What did you learn this week that should be in your prompt?"

A7 · The AI Onboarding Agent — "Newton" (Month-4 Deliverable)

By month 4, three businesses are running on agent stacks. The patterns are clear. Newton performs the interview + scaffolding that Jack did manually three times — in a fraction of the time — so acquisition #5 in month 5 deploys an agent in week 1, not month 1.

In Scope (v1, Nov 20)

45-min structured interview (Discord + voice via Quo)
Map discovered tools against MCP catalog
Propose 6-10 candidate workflows from canonical library
Draft autonomy charter with trust thresholds
Output Notion runbook + n8n stubs + Discord channel structure
Generate "Week 1 Plan" with D→S transitions

Out of Scope (v1)

Building new MCP adapters (specs them only)
Configuring billing/payment flows (always human)
Spinning up new infrastructure (compose/Portainer/DNS)
Model choice or routing decisions

Success criteria: Interview-to-runbook <4h · Proposed workflows accepted by Jack ≥70% · Tool catalog accuracy: zero false in-catalog flags · Newton-to-deployed-agent <14 days · Acquisition #5 reaches V tier on ≥3 workflows by end of Nov.

TRACK B

Acquire SMB #5 by Oct 15 (and Line Up #6)

$1.5M-$3M revenue, $150-500k SDE, labor >25% of revenue. Pay 3.0-3.5x stated SDE, achieve ≤2.5x post-AI. Sellers can't price in the labor compression. Brokers can't underwrite it. That's the alpha.

B1 · Target Box

Hard Gates

Revenue $1.5M-$3M TTM (audited or reviewed)
SDE $150-500k; labor cost ≥25% of revenue (binding)
Asking multiple ≤3.5x stated SDE (walk-away above)
Customer concentration: top <20%, top-5 <50%
Revenue trend flat-to-growing TTM

Hard Disqualifiers

Revenue declining >5% TTM · Customer concentration >30%
Pending litigation >$50k · 1099 misclassification risk
Owner is >50% of revenue generation (key-person trap)
Pre-1995 tech stack with no API · Cash-basis books
Real estate entangled with operating business purchase

Industry Shortlist (ranked)

Rank	Vertical	Adjacency	Labor %	AI Automatability	Deal Flow
1	Additional STR portfolio (8-25 units)	Direct — FlexStay playbook	30-40%	60-70%	High
2	Property mgmt roll-up (LTR, 100-400 doors)	Adjacent — same MCP stack	40-50%	50-60%	Medium
3	Small bookkeeping/CAS firm (10-40 SMB clients)	Adjacent — Jack lives in QBO	55-65%	65-75%	Medium-High
4	Multi-location cleaning/lawn (3-8 crews)	Adjacent — vendor mgmt overlap	45-55%	35-45%	High
5	Boutique F&B (coffee/bakery, 2-4 locations)	Direct — Donuts playbook	35-45%	25-35%	Medium

Recommendation: Lead with #1 (STR portfolio — highest conviction) and #3 (bookkeeping — highest AI margin lift). #2/#4 secondary. Deprioritize F&B (Donuts concentration risk).

B2 · Pipeline KPI Tree

Stage	Weekly/Monthly Target	Conversion to Next	Time-in-Stage SLA
Targets sourced	25/week	40% pass screen	2 days
Initial screen pass	10/week	50% NDA	Same day
NDA signed	5/week · 20/mo	60% CIM reviewed	5 days
CIM reviewed	12/month	30% deep dive	3 days
Deep dive (mgmt call + 3yr P&L)	4/month	50% IOI	7 days
IOI sent	2/month	60% LOI	5 days
LOI sent	1.2/month	50% accepted	3 days
LOI accepted	0.6/month · 4 over sprint	50% to QoE	7 days
QoE / full diligence	2 over sprint	50% to close	30-45 days
Closed	1 by Oct 15	—	—

Funnel math: 25/wk × 22 wks = 550 sourced → 220 screen-passed → 110 NDAs → 66 CIMs → 22 deep dives → 11 IOIs → 6 LOIs → 3 accepted → 2 QoE → 1 closed. Bias toward over-sourcing in June.

B3 · Sourcing Channel Matrix

Channel	Est. Flow/Mo	Quality	Effort	Cost	Verdict
BizBuySell Pro	40-80 in box	Low (picked over)	2 hrs/wk	$60/mo	Use — saved searches
BizQuest / BusinessBroker.net	15-40	Low-Med	1 hr/wk	Free	Overflow
QuietLight	5-10 (e-com)	High	1 hr/wk	Free	Use for DTC
Local biz brokers (regional)	5-15 off-market	High	6-8 hrs to build 10 relationships	$0	Highest ROI
Direct owner outreach (LinkedIn/letter)	0-5/100 letters	Very High	8-10 hrs/wk	$200/mo Apollo + postage	Highest quality
Industry platforms (VRM Intel, Axial)	5-15	Med-High	2 hrs/wk	$0-200/mo	Use for verticals 1/2/5
Searcher communities (Searchfunder, Acquisitions Anonymous)	2-5 deal share	Med	1 hr/wk	Free	Use, low effort

Channel mix target (weekly): 40% broker-listed (volume), 35% direct outreach (quality), 25% off-market through owner communities.

B4 · Diligence Checklist — Red/Yellow/Green

Decision rule: ≥3 reds = walk. 5+ yellows = renegotiate price down 10-15% or walk.

Financial (8 items)

#	Item	Green	Yellow	Red
1	QoE report (3-yr)	Independent firm, no major adj	Owner-prepared	Refuses
2	Add-back verification	<10% of SDE, documented	10-20%	>20% or hand-wavy
3	Customer concentration	top <15%	15-25%	>25%
4	Revenue trend	+5% TTM	Flat	Declining
5	AR aging	<10% over 60 days	10-20%	>20%
6	Gross margin stability	<2pp 3-yr variance	2-5pp	>5pp
7	Working capital normalized	Net positive, predictable	Seasonal but stable	Chronically negative
8	Owner comp vs market	At market, documented	Above market (clean add-back)	Below market (hidden labor)

Operational (6 items)

#	Item	Green	Yellow	Red
9	SOPs documented	Written, current	Tribal but recoverable	Undocumented, owner-only
10	Key person dependency	Owner <20% of ops	20-40%	>40%
11	Employee tenure (avg)	>3 yrs	1-3 yrs	<1 yr
12	Tech stack modernity	Cloud, API-accessible	Hybrid	On-prem, closed
13	Vendor concentration	top <20%	20-35%	>35% or single-source critical
14	Maintenance capex	<3% rev	3-7%	>7% or unknown

Legal · Commercial · AI-Readiness (14 items)

#	Category	Item	Green	Yellow	Red
15	Legal	Contracts assignability	Assignable, written	Requires consent	Not assignable
16	Legal	Employee classification	Clean W-2	Mixed, documented	1099 with control
17	Legal	IP ownership	Registered, in entity	In entity, unregistered	Owner personal
18	Legal	Pending litigation	None	<$25k routine	>$50k or class action
19	Legal	Lease terms	3+ yrs, assignable	1-3 yrs	<1 yr or non-assignable
20	Commercial	Customer interviews (5-10)	80%+ satisfied, would stay	Mixed	Hostility surfaced
21	Commercial	NRR (if applicable)	>100%	90-100%	<90%
22	Commercial	Churn rate	<10% annual	10-20%	>20%
23	Commercial	Competitive position	Defensible niche	Undifferentiated	Losing share
24	Commercial	Pipeline / fwd bookings	60+ days visibility	30-60 days	<30 days
25	AI	Data accessibility	APIs, structured	Exportable	PDFs/paper
26	AI	Automatable workflow %	>40%	20-40%	<20%
27	AI	Existing software footprint	QBO, Notion, modern PMS	Industry-standard	Legacy/custom
28	AI	Staff change-readiness	Comfortable with tech	Trainable	Hostile to change

B5 · AI-Leveraged Underwriting Model

Purchase price = Stated SDE × Offered multiple
Labor savings = Labor cost × AI automation rate
Adjusted post-AI SDE = Stated SDE + Labor savings
Implied post-AI multiple = Purchase price / Adjusted SDE
Target: pay ≤3.0x stated, achieve ≤2.0x post-AI

Worked example — $1M deal at $250k SDE, $400k labor cost, 25% AI automation

Purchase price: $250k × 3.0x = $750k (negotiate from 3.5x ask)
Labor savings: $400k × 25% = $100k/yr
Adjusted SDE: $250k + $100k = $350k
Implied post-AI multiple: $750k / $350k = 2.14x ✓

Sensitivity Table (AI automation rate)

AI Savings	Adj SDE	Post-AI Multiple	Yr-1 CoC	5-yr IRR (sale @ 4.5x)
0% (bear)	$250k	3.0x	18%	22%
15%	$310k	2.42x	28%	31%
25% (base)	$350k	2.14x	35%	37%
30%	$370k	2.03x	39%	40%

Discipline: Always model 0% AI savings as bear case. If the deal doesn't clear 15% IRR with no AI lift, walk — you're betting the thesis, not buying margin of safety.

B6 · Financing Structure ($1M Deal Example)

Sources

Buyer equity (Jack)	$105,000	10%
Seller note	$200,000	19%
SBA 7(a)	$745,000	71%
Total Sources	$1,050,000	100%

Uses

Purchase price	$1,000,000
Closing costs (3%)	$30,000
Working capital reserve	$20,000
Total Uses	$1,050,000

Terms summary:

SBA 7(a): 10-yr amort, Prime+2.75% (~10.75% all-in 2026), 3% SBA fee financed, personal guarantee
Seller note: 5-yr, 6%, standby yrs 1-2 (debt service starts yr 3)
Annual debt service: $120k yrs 1-2, $166k yrs 3-5 · DSCR ~2.1x at post-AI SDE

B7 · Month-by-Month Execution

Month	Focus	Targets	Exit Criteria
Jun	Foundation	Lock target box · SBA pre-qual letter · 10-15 broker relationships · Deploy Scout + Notion DB + underwriting calc	20 NDAs signed · 40+ targets sourced
Jul	Funnel fill	Review 12 CIMs · 4 mgmt calls · Send 2 IOIs · Eliminate underperforming verticals	6-8 active deep-dive targets
Aug	First LOI	2-3 deep diligence reviews · Customer ref calls · Engage QoE selectively	First LOI sent by Aug 15
Sep	Diligence + financing	Full QoE, legal, AI-readiness audit · SBA underwriting in parallel · Backup deal in late-stage	Cleared QoE on primary target
Oct	Close	Day-1 integration playbook: Maple/Mickey MCP → acquired QBO/Gmail · SOPs ingested · AI savings stopwatch starts	Close by Oct 15
Nov	Prove + start #6	30-day post-close labor cost measurement (target: -15% by day 30, -25% by day 90) · Publish case study	LOI on acquisition #6 by Nov 15

B8 · Tracking Tools — BUILD vs USE

USE (off-the-shelf)

BizBuySell Pro ($60/mo) — saved searches, alerts
Apollo.io ($99/mo) — owner contact enrichment
DocSend ($45/mo) — teaser/CIM sharing with view analytics
Crunchbase Pro ($684/yr) — competitive context
Plaid / Codat (per-deal) — bank/QBO pull for QoE
DealRoom or Midaxo (~$500/mo) — only at LOI stage

BUILD (Jack-specific)

Notion "Track B Deal Pipeline" DB — full schema with auto-calc multiples and Discord status pings
AI-Leveraged Underwriting Calculator — Next.js page at underwriting.jackyen.com, ~4 hrs build
Scout — weekly deal review agent — Mon/Wed/Fri 7am cron · scrapes BizBuySell + BizQuest · scores against red flags · auto-inserts to Notion · posts to Discord #track-b-deals · ~2 days build

TRACK C

Content Engine — Distribute the Exhaust

Every action in Track A and Track B generates raw material. Track C is a near-zero-marginal-cost distribution layer that compounds that material into deal flow, podcast invites, and optional consulting income. The content is the exhaust, not the engine.

C1 · Audience Funnel KPI Tree

North Star

Qualified Inbound per Month — DMs/emails that fall into (a) deal-flow leads, (b) consulting inquiries with budget signal, or (c) podcast invites with >5k audience.

Target: 0 → 8-20/mo by month 6.

L1 — Reach

X / LinkedIn followers
Newsletter subs
Podcast guest spots (cumulative)
Monthly impressions

L2 — Engagement

Engagement rate (3%+ target)
Profile visits per post
Reply rate · Newsletter open rate (40%+)
CTR (5%+)

L3 — Operations

Posts per week (by channel)
Format mix (thread / single / image / screenshot)
Topic resonance score per pillar

C2 · Channel Allocation

70%

Twitter / X

Operator-builders + search funders live here. Velocity tolerates 10+ posts/week. Algorithm rewards consistency over polish.

20%

LinkedIn

Lower velocity, higher per-impression deal-flow value. SMB sellers, brokers, family offices read LinkedIn.

10%

Substack

Email is the only channel without algorithmic permission gates. Bi-weekly cadence. Compounding moat.

Explicit exclusions: YouTube long-form (4-8h per video), TikTok/Reels (wrong demo), Reddit (low compounding), Threads/Bluesky (too small — revisit Q1 2027).

C3 · Content Pillars

#	Pillar	Description	Cadence	Example Hook
1	"What Maple/Mickey/Goldie did this week"	Concrete action + screenshot + anonymized $/hours	2× X / 1× LI per week	"Mickey processed 47 guest messages overnight. Here's the one I had to override and why."
2	Ops Cost Teardowns	Before/after metrics — minutes, dollars, error rate	2× X / week	"Hostfully → Notion → Discord approval used to take 8 min. Now: 90 seconds. Here's the wiring."
3	Acquisition Diaries	Track B journey, no deal specifics until LOI	1× X / 1× LI per week	"Week 4 of SMB #5 search. Reviewed 11 CIMs. Killed 9 in <2 minutes each. Here's my filter."
4	Build Logs	Technical — MCP, n8n, agent architecture	2× X / 1 essay per month	"Why I moved my agent's memory from JSON to a 4-layer Obsidian vault."
5	Operator Interviews	5-Q Q&A with other operator-builders	1× every 2 wks (ramps mo3+)	"5 questions with [operator] on running a 7-figure laundromat with 1 employee + AI."

C4 · Weekly Cadence

Day	Twitter/X	LinkedIn	Substack
Mon	8am Build Log · 12pm Agent-did-this	—	—
Tue	8am Ops Teardown · 12pm Agent-did-this	8am Acquisition Diary	—
Wed	8am Build Log · 12pm Ops Teardown	—	—
Thu	8am Acquisition Diary · 12pm Agent-did-this	8am Ops Teardown	—
Fri	Reply day (30 min) · 12pm Agent-did-this	—	—
Sat	9am Light post/opinion	8am Build Log / Operator Q&A	—
Sun	9am Weekly Rollup Thread	—	7am (every other) Essay

Floor: 10 X / 3 LI / 1 Substack every 14 days / 1 Sunday rollup thread per week.

C5 · The Sunday Rollup — Signature Artifact

Format: Twitter thread (8-14 tweets) + cross-posted to LinkedIn (single long-form) + summarized in Substack every other week.

Fixed Structure (8 sections)

Headline number — total hrs saved by agents this week
Agent box score — Maple / Mickey / Goldie actions, override rate
$$ moved — anonymized dollar throughput
Acquisition pipeline — CIMs reviewed, calls taken, stage of furthest deal (no names)
Wins — 2-3 concrete shipped things
L's / lessons — 1-2 things that broke + what changed
Next week's bet — public commitment for accountability
Open question — invites replies, fuels engagement

Assembly Pipeline (week 1 build)

Fri 5pm: RollupBot agent pulls from:
- QBO classes (FlexStay/GG/Donuts/Property) for $$
- Notion msg DB 355671dd-92cc-80d7-9dd5-000b34836b53
- Discord channels (1503028624096366792 daily rollup)
- Notion "Acquisition Pipeline" DB
Drafts thread in markdown → #content-drafts
Sat morning: Jack edits, sanitizes, schedules via Tweet Hunter for Sun 9am

C6 · Podcast Pitch Sequence

#	Show	Angle	Target Month
1	Acquiring Minds (Will Smith)	"Searching for SMB #5 while running 4 with AI agents"	Jul
2	Owned and Operated (Connor Groce)	Multi-location ops with one operator + agent team	Aug
3	Sweaty Startup (Nick Huber)	Unsexy holdco angle, FlexStay ops	Aug
4	SMB Attorney Podcast (Eric Pacifici)	Operator-buyer credibility post-LOI	Sep
5	My First Million	"AI-leveraged holdco" thesis as standalone idea	Sep
6	Get Paid For Your Pad (Jasper Ribbers)	FlexStay STR + Hostfully automation	Sep
7	The Bootstrapped Founder (Arvid Kahl)	Solo-operator-with-AI as 1-person leverage	Oct
8	In The Pit (Khe Hy)	Productivity at the holdco-operator level	Oct
9	Hampton (peer founder pod)	Peer-network credibility	Oct
10	The Hustle Daily Show	News-hook: "guy runs 4 SMBs with AI"	Oct
11	Capital Stack (Paul O'Brien)	Search fund / capital allocation framing	Nov
12	Founders Podcast (David Senra)	Operator-as-craftsman; long shot	Nov

Pitching rule: Don't pitch until month 2 — need 3-4 weekly rollups as proof of cadence first.

C7 · Six-Month Growth Targets

Month	X Followers	LI Followers	Newsletter	Podcast Apps (cumulative)	Inbound Qualified/mo
Jun (M1)	900	1,800	80	0	1
Jul (M2)	1,600	2,200	200	1	2
Aug (M3)	2,800	2,800	400	3	4
Sep (M4)	4,500	3,500	750	5	8
Oct (M5)	7,500	4,200	1,200	7	13
Nov (M6)	10,000-15,000	5,000	1,500-2,000	6-8	15-20

Base case (X = 7-10k) is achievable on cadence alone. Stretch case requires 1 viral hit (>500k impressions).

C8 · Inbound Triage

Cat	Looks like	Owner	SLA	Conversion target
A. Deal flow	Broker, seller, intermediary	Jack personally	24h	30% → discovery call
B. Consulting inquiry	"Can you help me set this up?"	Agent drafts, Jack approves	72h	15% → paid discovery
C. Podcast invite	Host or producer	Jack personally	24h	80% → booked if fit
D. Fan / curious	"Cool what you're building"	Agent auto-reply (warm)	72h	Log only
E. Peer operator	Wants to commiserate	Agent drafts intro	72h	Build peer network

C10 · Implementation Steps — Month by Month

Concrete steps mirroring the cadence in C4 and the artifacts in C5-C8. Each item is a measurable thing Jack (or his agent) ships in that month. If a step slips two consecutive weeks, treat it the same as a Track A/B red flag — pause and triage.

Week 1 (Jun 1-7) · Foundation

Open Substack account, reserve handle (jackyen.substack.com or similar), draft welcome auto-email.
Open / verify Tweet Hunter account; set up scheduling slots for the 7-day cadence in §C4.
Stand up Content Performance Notion DB (B14) with schema from §C7 (post_url, channel, date, format, pillar, hook_template, impressions, engagement_rate, replies, profile_visits, inbound_DMs, outcome, notes).
Stand up Hook Library Notion DB (B16) with schema (hook_text, pillar, first_used_date, best_post_url, avg_impressions, times_reused, retire_after_date). Seed with 10-15 starter hooks from §C3 examples.
Draft Sunday Rollup template v1 — the 8-section structure from §C5, saved as a Notion page template.
Ship 3 X posts + 1 LI post this week — one per pillar (Build Log + Agent-did-this + Ops Teardown). Voice/topic calibration.
Saturday morning: hand-assemble + schedule the very first Sunday Rollup for Sun 9am.

Weeks 2-4 (Jun 8-30) · Rhythm

Hold the cadence: 10 X / 3 LI / 1 Sunday Rollup per week — zero skips. If traveling, schedule Friday.
Ship first Substack essay (2,000-3,000 words). Topic: pillar #4 (Build Log) — e.g., "Why I run my AI coworkers in Discord, not Slack."
Every Friday: log the top 3 posts of the week into the Hook Library DB (pillar tagged, impressions noted).
Every Sunday during the 90-min ops review: skim Content Performance DB → adjust next-week pillar mix (drop any pillar <50% baseline).
Reach 1,000 X followers by Jun 30 (from ~600 baseline). If still <800 by Jun 24, audit hook quality with a small sample of replies.
Reply day discipline: Friday 30-min block in replies to operator-builders / search funders / SMB content. No bot replies.

Month 2 (Jul) · Pitch + RollupBot

Build RollupBot (B15) — n8n Friday-5pm cron that pulls QBO classes + agent_action_log + Track B Notion DB + Discord daily-rollup, drafts the 8-section thread to #content-drafts. ~8 hours of build effort.
Pitch podcasts #1 (Acquiring Minds) and #2 (Owned and Operated) using the §C6 template. Need 3-4 Sunday Rollups already shipped as proof of cadence.
Add "Acquisition Diary" pillar to the rotation now that Track B sourcing is active (CIMs reviewed = real raw material). Tag posts with acq-diary.
Start the Inbound DM Triage agent (B17) — watches X/LI DMs and Gmail, categorizes A-E per §C8, posts to Discord #inbound-triage.
End-Jul Substack essay: pillar #2 (Ops Teardown) — first 60-day savings numbers from FlexStay or Goldenglaze.
Target: 1,600 X / 2,200 LI / 200 newsletter subs / 1 booked podcast.

Month 3 (Aug) · Compound + first case study

Ship the first public case study post — anonymized 60-day numbers across the portfolio (hours saved, dollars moved, override rate). Cross-post to LI as a single long-form, push as Substack essay.
Pitch podcasts #3 (Sweaty Startup) and #4 (SMB Attorney). Book at least 1 for an Aug/Sep recording.
Run the first Hook Library refresh routine: retire any hooks <2× baseline impressions over 30 rolling days. Clone the top 3 hooks into 3 variants each.
Publish the first Operator Interview (pillar #5) — 5-Q Q&A with another operator-builder Jack respects.
Run "RollupBot dry test" — let it draft 2 weeks of Sunday threads, Jack edits and ships. Measure draft acceptance >70%.
Target: 2,800 X / 2,800 LI / 400 newsletter / 3 cumulative podcast bookings / 4 qualified inbound DMs/mo.

Month 4 (Sep) · Inbound at scale

Inbound DM Triage agent fully live — A (deal flow) and C (podcast) escalate to Jack; B/D/E drafted by agent.
Pitch podcasts #5 (MFM) and #6 (Get Paid For Your Pad). Record at least 2 episodes this month.
Acquisition Diary content peaks (Track B LOI window). Threads about "what I look for in CIMs" + "the deal I killed at last week's call" without specifics.
Mid-month Substack: "The AI-leveraged underwriting model" — share the equations + sensitivity table from §B5. This is the post that gets shared in private SMB acquirer Slacks.
Target: 4,500 X / 3,500 LI / 750 newsletter / 5 cumulative podcasts / 8 qualified inbound/mo.

Month 5 (Oct) · Close + flagship case study

Day-of-close: post a single screenshotted Discord moment ("just signed for SMB #5"). No specifics, just the screenshot.
Day-14 post: "How we got AI-leveraged ops live in 14 days" — anonymized walk-through. This becomes the flagship case study for ai-business.jackyen.com.
Ramp newsletter cadence to weekly (was bi-weekly) — the acquisition story has enough material per week.
Pitch podcasts #7-10. Book 2-3 for October.
Triage first paid consulting inquiry — only accept if Jack has ≥5 qualified DMs that week (proves pipeline). Quote $35-50k for a 90-day done-with-you engagement.
Target: 7,500 X / 4,200 LI / 1,200 newsletter / 7 cumulative podcasts / 13 inbound/mo.

Month 6 (Nov) · Maturity + the 6-month rollup

Mid-month: "30-day proof" post — labor cost reduction at SMB #5, charts.
Pitch podcasts #11 (Capital Stack) and #12 (Founders Podcast). Book whatever lands.
The 6-Month Rollup Substack essay (Nov 17ish) — definitive long-form covering: hours saved, businesses owned, acquisitions in pipeline, lessons. This is the artifact you reference for the next 24 months.
Audience review: which pillar performed, which retired. Decide 2027 mix.
Decision: keep 10/wk X cadence OR invest the freed time into YouTube long-form (one 8-min essay/month). Default = stay on X.
Target: 10,000-15,000 X / 5,000 LI / 1,500-2,000 newsletter / 6-8 cumulative podcasts / 15-20 inbound/mo.

The single non-negotiable: the Sunday Rollup ships every week for 26 consecutive weeks. Missing one breaks the build-in-public moat. If the data isn't there to fill 8 sections, that's a Track A/B problem, not a content problem.

C9 · What NOT to Do

No generic AI hot takes ("GPT-5 changes everything")
No engagement-bait threads ("9 mental models that 10x'd my business")
No paid amplification in months 1-3 — organic-only proves the format
No "build your own AI agents" tutorials — show what, not full how
No financial details that affect deals in flight; until LOI, industry+size band only
No subtweeting, politics, or public vendor fights
No ghostwriting — Jack writes or edits every post; the voice is the moat
No skipping Sunday Rollup — if traveling, schedule it Friday

FINANCE

Financial Framework — KPI Tree, Baselines, Cash

Every number Jack touches in the next 6 months should ladder up to the North Star. If it doesn't, kill it.

F1 · The KPI Tree

NORTH STAR — Annualized Post-AI EBITDA across all owned businesses
Target: $850K (May 20) → $1.35M (Nov 20) → $2.0M (May 2027 w/ SMB #5)

├── L1.A — Portfolio Revenue ($/yr, run-rate)
│   ├── L2.A1 — Revenue per business (FlexStay / Donuts / Property / GG)
│   │   ├── L3 — Booked nights / occupancy % (FlexStay)
│   │   ├── L3 — Units/day × avg ticket (Donuts)
│   │   ├── L3 — Occupied units × avg rent (Property)
│   │   └── L3 — Orders/mo × AOV (Goldenglaze)
│   ├── L2.A2 — Customer concentration (top 5 % of rev)
│   └── L2.A3 — Revenue retention / churn
│
├── L1.B — Portfolio Ops Cost ($/yr)
│   ├── L2.B1 — Labor cost ($/mo per business)
│   │   ├── L3 — Labor hrs/wk per business
│   │   ├── L3 — Fully-loaded $/hr by role
│   │   └── L3 — Headcount (FTE + 1099)
│   ├── L2.B2 — Non-personnel ops cost ($/mo per business)
│   ├── L2.B3 — Model & tooling spend ($/mo)
│   │   ├── L3 — Anthropic/OpenAI tokens/mo
│   │   ├── L3 — MCP infra (n8n, Notion, Discord, hosting)
│   │   └── L3 — $/agent action
│   └── L2.B4 — Error remediation cost ($/mo — refunds, comp nights, redo)
│
├── L1.C — AI-Driven Savings Rate (% of pre-AI ops cost eliminated)
│   ├── L2.C1 — Hours saved/wk per business
│   ├── L2.C2 — Automation coverage % (tasks automated ÷ tasks catalogued)
│   ├── L2.C3 — Agent-handled actions/mo (volume)
│   │   ├── L3 — # approved-as-drafted (no human edit)
│   │   ├── L3 — # edited-then-approved
│   │   ├── L3 — # escalated to Jack
│   │   └── L3 — # rejected / unsafe
│   └── L2.C4 — Net savings $/mo (labor saved − model spend − error cost)
│
├── L1.D — # of Businesses Owned (count)
│   ├── L2.D1 — Active business count (currently 4 → 5 by Oct)
│   ├── L2.D2 — Pipeline (LOIs out, IOIs received)
│   └── L2.D3 — Days-to-AI-live on new acquisitions
│       ├── L3 — SOPs encoded per business
│       ├── L3 — MCP integrations live per business
│       └── L3 — Agent on-boarded to Discord (Y/N + date)
│
└── L1.E — Acquisition Multiple Expansion (turns of EBITDA created)
    ├── L2.E1 — Pre-AI SDE multiple paid (target ≤3.5x)
    ├── L2.E2 — Post-AI SDE multiple effective (target ≤2.5x)
    ├── L2.E3 — Implied IRR on deployed equity (target ≥35%)
    └── L2.E4 — Holdco-level EBITDA multiple if marketed (target ≥5.5x)

Read it this way: if Net savings $/mo (L2.C4) shrinks while model spend (L2.B3) grows, the thesis is bleeding. If Automation coverage % (L2.C2) stalls below 60% by month 4, productization is failing and acquisition #5 is premature.

F2 · Per-Business Baseline + 6-Month Targets

Business	Current Labor Hrs/Wk (est)	Current Ops Cost (% rev)	6-Mo Target Hrs/Wk	6-Mo Cost Reduction	QBO Class / Location	QBO Reports + Frequency
Goldenglaze P1	8 (CS + fulfillment ops)	22%	4	−$650/mo + −15% CS response time	Class: Goldenglaze	P&L by Class monthly · Sales by Customer weekly · Stripe recon weekly
FlexStay P2	40 (Jack + 1 VA + cleaners coord)	38%	18	−$4,200/mo labor · −$800/mo refunds	Class: FlexStay · sub-Loc per cluster	P&L by Class monthly day 5 · Payroll Detail biweekly Mon · A/R Aging weekly Mon · Vendor by Class monthly
Donuts P3	60 (2 bakers + 1 counter + Jack)	52%	35 (automate ordering/sched/inv)	−$2,100/mo shrink + scheduling waste	Class: Donuts	P&L by Class monthly · Inventory Valuation weekly Sun · Payroll Detail weekly Mon · COGS % weekly trend
Property P4	12 (rent collection + maint coord)	18%	6	−$900/mo · +$1,400/mo faster turn	Class: Property · Location per address	P&L by Location monthly · Tenant A/R weekly Mon · Maintenance spend monthly
PORTFOLIO	120 hrs/wk	~36% blended	63 hrs/wk (−47%)	−$7,850/mo = −$94K/yr	Consolidated	Consolidated P&L by Class monthly day 5 · Aged Payables weekly · Cash Flow Forecast weekly Wed

MCP automation: All QBO pulls run via mcp__quickbooks__* calls in an n8n workflow on Wed 6am cron → writes to Notion Portfolio Ops Metrics → posts summary to Discord #weekly-review.

F3 · Personal Financial KPIs — Owner Cash Flow

Metric	Target	Cadence	Source
Owner draw — FlexStay	$8K/mo (Aug) → $12K/mo (Nov)	Monthly	QBO Equity > Owner Draw, Class=FlexStay
Owner draw — Donuts	$4K/mo steady	Monthly	QBO Class=Donuts
Owner draw — Property	$3K/mo	Monthly	QBO Class=Property
Owner draw — Goldenglaze	$1.5K/mo · reinvest rest	Monthly	QBO Class=Goldenglaze
Total monthly owner cash	$16.5K → $20.5K by Nov	Monthly	Consolidated
Personal liquidity reserve	≥$150K cash + $50K LOC undrawn	Weekly snapshot	Personal banking
Acquisition equity reserve	$200K-$300K segregated by Sep 1	Monthly	Separate HYSA
Personal burn rate	≤$11K/mo	Monthly	Mint/Copilot
Cash runway w/o distributions	≥9 months	Monthly	Reserve ÷ burn

SBA 7(a) Pre-Qual Checklist (start now — ~90 days to deal-ready)

☐ 3 yrs personal tax returns clean, filed, signed
☐ 3 yrs business returns for FlexStay (highest-EBITDA, primary collateral)
☐ Personal Financial Statement (SBA Form 413) drafted, updated quarterly
☐ Personal credit score ≥720 (pull tri-merge; freeze inquiries)
☐ DSCR ≥1.5x demonstrable on target #5 pro forma
☐ 10% equity injection liquid and seasoned ≥60 days in account
☐ Industry experience narrative (the holdco-operator story)
☐ Pre-qualified with 2-3 SBA Preferred Lenders by Aug 1
☐ Business valuation reports framework lined up

F4 · Red Flags — When the Thesis Fails (2-week breach = pause)

#	Red Flag	Trigger	Why It Kills the Thesis
1	Model spend / labor savings ratio	>25% (target <15%)	AI more expensive than humans it replaces
2	Agent error rate	>3% of actions, OR single error >$2K loss	Trust collapses; rollout must slow
3	Escalation rate trending up	>20%, WoW growth	Agents aren't learning; Jack = bottleneck
4	Customer concentration	Single customer >25% of any business's rev	One churn destroys leveraged EBITDA
5	FlexStay owner churn	>2 off-board in 90 days	Inventory-loss; can't replace fast
6	Personal liquidity	<$100K cash for >30 days	Forced sale risk; #5 dies
7	Days-to-AI-live slipping	Any business <60% automation by Sep 20	Productization failing; don't add #5
8	Acquisition multiple creep	Any LOI >3.5x pre-AI SDE	Math gets thin even with AI
9	Content cannibalizing ops	Content time >6 hrs/wk for Jack	Flywheel reversed; track C is byproduct, not job
10	EBITDA expansion below 10% on any productized biz	By month 4	That business may be fundamentally non-AI-leverageable

TOOLS

Consolidated Tool Stack — BUILD vs USE

~114 hrs of BUILD effort over 6 months (~5 hrs/week sustained). Everything else USES tools Jack already pays for.

BUILD (commissioned work)

#	Tool	Track	Hours
B1	Agent Action Log Notion sync (n8n)	A	6
B2	Postgres analytics mirror	A	8
B3	Weekly rollup poster (n8n, Mon 7am)	A	6
B4	Charter card automation	A	10
B5	POS → Drive watcher (Donuts)	A	4
B6	Commerce MCP for Goldenglaze	A	16
B7	OCR pipeline for vendor invoices	A	12
B8	Trust-threshold guardrail middleware	A	8
B9	Newton — AI Onboarding Agent v1	A	40
B10	Property tenant portal email parser	A	4
B11	Track B Deal Pipeline Notion DB	B	3
B12	AI-Leveraged Underwriting Calculator (Next.js)	B	4
B13	Scout — Acquisition sourcing agent	B	16
B14	Content Performance Notion DB + Twitter sync	C	3
B15	RollupBot — Sunday thread drafter	C	8
B16	Hook Library Notion DB	C	2
B17	Inbound DM triage agent	C	6
	Total BUILD		~156 hrs

USE (existing / off-the-shelf)

Tool	Usage
QBO MCP	Recon (F3, D1, D6, P3, G5), reporting (F4, P5), AR aging
Hostfully MCP	F1-F2, F5-F7 — all guest/cleaner workflows
flexstay-ops MCP	F2, F5, P2 — job dispatch shared FlexStay+Property
Gmail MCP (7 accounts)	Triage across all 4 businesses; filter creation
Notion MCP	Action log, charter cards, case studies, FAQs, tenant/owner DBs
n8n	Glue — every cron, every webhook, every rollup
Quo MCP	Outbound SMS, inbound triage for Property
Discord (Maple/Mickey/Goldie)	Human review cockpit (never source of truth)
Google Calendar MCP	P4 lease renewals, F2 cleaner scheduling
BizBuySell Pro	$60/mo · saved searches + alerts
Apollo.io	$99/mo · owner enrichment for direct outreach
DocSend	$45/mo · CIM/teaser sharing with analytics
Tweet Hunter	X scheduling + hook library
Substack	Newsletter (migrate to Beehiiv if subs >2k)
Plaid / Codat	Per-deal QoE bank/QBO pull

SKIP (for now)

Mosaic / Jirav / Fathom — revisit Q1 2027 when 5+ entities + board
BuzzSumo — too expensive for current scale
YouTube long-form, TikTok, Reddit, Threads/Bluesky
Paid amplification (revisit month 4+)

Observability Gaps (today)

Gap	Why it exists	Closure
Agent stdout is ephemeral	No structured log emitter in OpenClaw runtime	B1 writes structured records at every tool call → stdout becomes greppable in Postgres
No "agent vs human" attribution in outcomes	Hostfully tags by user account only; agent posts under Jack's account	Hidden tag (e.g., [ai-rev:WF-F1-v3]) appended; n8n computes per-agent CSAT
Cleaner-completion truth is self-reported	set_job_status is a one-button flag	Require photo upload via Quo SMS reply → bind to job via MMS → Drive → job thread
POS data lag (Donuts) — intraday blind	Square/Toast APIs not wired	Defer to mo-3 unless intraday decisions matter
No "would have approved" ground truth in Shadow tier	No interface to triage shadow drafts at volume	Lightweight Notion view: cards with Approve/Reject/Edit buttons (part of B1)

Consolidated Risk Register

#	Risk	Track	Prob	Impact	Mitigation
1	SBA financing falls through	B	Med	High	Pre-qual in Jun · backup lender (Live Oak, Newtek) · seller carry 25% as fallback
2	Seller misrepresents financials	B	High	High	Independent QoE non-negotiable · bank stmts > P&L · tax returns > both
3	Key employee departs at announcement	B	Med	High	Stay bonuses 10-20% salary, 12-mo vest · identify in diligence
4	Customer concentration revealed in diligence	B	Med	High	Top-10 disclosure required at LOI · price adjustment for late discoveries
5	AI savings thesis under-delivers (<10%)	A/B	Med	Med	Underwrite to 0% AI case · only deploy on deals that work at 3.0x stated
6	Timeline slips past Oct 15	B	High	Med	Run 2 deals in parallel from Aug · ok to close Nov-Dec for quality
7	Cultural integration failure (staff rejects AI)	A/B	Med	High	AI-readiness scored in diligence (item #28) · first 90 days = augment not replace
8	Deal fatigue / opportunity cost on existing 4 SMBs	A/B	High	Med	Track B time-boxed to 15 hrs/wk · Scout/Maple handle sourcing + screen autonomously
9	Anthropic / model price spike (+30%)	A	Med	Med	Multi-model routing (Haiku/Sonnet fallback) · quarterly price review
10	Agent error / churn event	A	Med	High	Mandatory HITL for first 60 days · trust thresholds enforced by middleware
11	Implementation labor bottleneck (Jack the bottleneck)	A	High	High	Newton (B9) is the unlock · prioritize ruthlessly · contractor for B5/B7
12	Content track creep — Jack becomes writer instead of operator	C	Med	Med	RollupBot drafts · Jack edits only · hard cap 6 hrs/wk on content

Week 1 Next Actions (May 20-26)

TRACK A

Baseline labor hours across all 4 businesses (7-day timekeeping)
Build Agent Action Log Notion DB (B1) + n8n hourly sync
Stand up Goldie in Goldenglaze Discord; deploy G1, G3, G4, G5 in Shadow tier
Audit existing Maple/Mickey/Goldie coverage; tag stale workflows
Confirm portainer compose entry for any new n8n workflows

TRACK B

Pull tri-merge credit report; SBA Form 413 draft started
Open relationships with 5 local biz brokers (Mon-Wed calls)
Set up BizBuySell Pro saved searches matching target box
Stand up Track B Deal Pipeline Notion DB (B11)
Email 3 SBA Preferred Lenders for intro call

TRACK C

Lock posting cadence in Tweet Hunter (10/wk X, 3/wk LI)
Ship 2 X posts + 1 LI post this week — voice + topic test
Draft first Sunday Rollup template (8-section)
Stand up Content Performance Notion DB (B14)
Open Substack account · reserve handle

DECISION RULE

If by Jul 31 fewer than 15 CIMs reviewed → sourcing is broken; escalate to direct outreach.
If no LOI sent by Aug 31 → quality bar too narrow; recalibrate but don't lower discipline.
If first deal closes but 90-day AI savings <10% → variant view is wrong; pause #6, focus on operational proof.
If SBA appetite tightens beyond P+3.5% → restructure with higher seller note (30%+) or delay close.