rewritten 2026-05-20

two operating businesses. one internal AI operating system.

Jack owns and operates Golden Glaze Donuts and FlexStay Housing. The AI plan is not a separate agency, not a SaaS company, and not four different business lines. It is an internal operating system that lowers labor load, tightens controls, supports the planned $1M Golden Glaze raise, and makes both businesses easier to scale.

the thesis

AI should make the operating companies stronger before it becomes anything external.

The real asset is not a generic AI agency. It is Jack's ability to install AI into messy, local, operational businesses where the work is repetitive, time-sensitive, and currently handled through people, texts, spreadsheets, Discord, PMS tools, POS tools, and bookkeeping cleanup.

Golden Glaze Donuts is the QSR proving ground, with the next unlock being a planned $1M raise and centralized kitchen. FlexStay Housing is the hospitality/property proving ground. If the same agent stack can reduce manual work in both, the playbook becomes credible for future acquisitions or services later. For now, the plan stays focused on the two companies Jack actually has.

do not count as businesses

Golden Ops is internal software for Golden Glaze, not a separate operating business.
Selling premium is a personal capital strategy, not an operating company in this plan.
Property should not be split out from FlexStay unless Jack says a separate long-term rental company exists.
Donut means Golden Glaze Donuts.

operating base

the two businesses

Golden Glaze Donuts

Golden Glaze = Donut = Golden Glaze Donuts. Same business.

DFW shops today

20+

employees

shop target by year-end

$1M

planned raise

central

kitchen plan

506(b)

investor structure

AI use cases

daily Discord check-ins, issue routing, and missed-check-in detection
cash control, deposit variance, register/deposit bag mismatch review
inventory and low-stock alerts tied to shop operations
central kitchen production planning, transfer controls, and waste visibility
labor/overtime visibility once payroll data is wired cleanly
investor/owner reporting support without spinning weak months

FlexStay Housing

STR/MTR housing across DFW, Hawaii, and Tampa.

units

markets

24/7

guest ops pressure

ops

owner + vendor coordination

AI use cases

guest message classification and drafted replies for Jack approval
maintenance triage, severity routing, and vendor follow-up
daily rollups for check-ins, check-outs, vacancies, and open issues
case-manager, owner, and vendor drafts separated from guest-side alerts
timezone-aware property operations across DFW, Hawaii, and Tampa

golden glaze capital plan

$1M raise + centralized kitchen

The Golden Glaze expansion path is not just more stores. It needs a centralized kitchen that can standardize production, protect quality, reduce store-level prep burden, and give the 7-to-10 shop plan a real operating backbone.

AI's role is to make the raise and kitchen build more bankable: clean operating data, tighter cash controls, repeatable production metrics, investor reporting support, and early warning when store execution slips.

Workstream	What needs to happen	AI assist
raise prep	Update investor deck, use-of-funds, store economics, central kitchen case, risk factors, and data room.	Generate diligence packet drafts from verified source data; flag missing or stale numbers.
kitchen model	Define facility needs, equipment list, production capacity, staffing, routes, transfer process, and opening budget.	Build capacity model, daily production planner, purchasing forecast, and waste tracker.
store rollout	Decide which products move to central production first and how each shop receives, stores, and reports usage.	Track transfers, spoilage, stockouts, variance, and shop-level compliance.
controls	Separate raise proceeds, capex, operating spend, and investor reporting cadence.	Weekly controller packet with cash, capex, variance, and milestone status for Jack review.

operator os

the AI stack is shared infrastructure.

Maple is the main operator interface. Mickey/Hermes handles durable backend jobs. Goldie is the Golden Glaze operating persona. The point is not to make more bots. The point is to turn recurring work into reliable workflows with approvals, logs, and escalation rules.

Layer	What it does	Current focus
Maple	Jack-facing operator. Routes work, remembers corrections, coordinates pages, notes, and subagents.	Keep business facts clean and prevent cross-brand drift.
Mickey / Hermes	Durable task runner for crons, backend workflows, and ops rollups.	FlexStay daily rollup, controller cash/variance pilot.
Goldie	Golden Glaze ops persona for store operations and Discord workflows.	Daily check-ins, store exceptions, cash controls, labor visibility.
Golden Ops	Internal software Jack built to run Golden Glaze operations.	Treat as infrastructure inside Golden Glaze, not a standalone company.

6-month plan

what has to happen next

1. clean the operating data

The AI layer is only as good as the inputs. Golden Glaze needs clean POS, cash, payroll, issue, and QBO visibility. FlexStay needs clean PMS, message, maintenance, calendar, and finance signals.

resolve QBO account/entity mapping issues
wire Golden Glaze payroll rows and variance denominators
keep FlexStay property IDs, markets, and timezones normalized

2. automate the repeat work

Push the most repetitive workflows into supervised automation first. Do not chase flashy AI demos. Automate what Jack, managers, vendors, and guests repeatedly force through text and spreadsheets.

store check-in watch and exception summaries
guest issue triage and drafted replies
cash/variance review and weekly controller packet

3. turn proof into leverage

Once the system proves it saves time and catches misses, use it to scale the two businesses. Future acquisition or agency work should come after proof, not before it.

show before/after hours saved by workflow
track avoided misses: late replies, cash exceptions, open issues
document the operator OS as Jack's public proof

Phase	Golden Glaze	FlexStay	Decision gate
May-Jun	cash variance, issue watch, check-in visibility, $1M raise packet outline	daily rollup, guest triage, maintenance routing	is the data clean enough to trust weekly reporting?
Jul-Aug	inventory, labor, vendor invoice workflows, central kitchen model	vendor follow-up, owner/case-manager drafts, exception alerts	are managers/Jack approving drafts without heavy rewrites?
Sep-Oct	multi-shop growth controls for 7 → 10 shops, kitchen launch controls	market-level operating controls for 11+ units	does the system reduce Jack's direct operating load?
Nov	owner/investor reporting pack, raise/kitchen milestone review	owner/ops performance pack	is this repeatable enough to publish or apply to the next deal?

execution planner

day-by-day plan

This is the operating checklist. Every day should push one of three outcomes: cleaner data, less manual work, or proof that the AI operating system is helping Golden Glaze Donuts and FlexStay Housing run tighter.

default daily rhythm

7:00 ops review · 8:00-11:00 build block · 1:00-3:00 operating calls/vendor follow-up · 4:00 scorecard update · 5:00 content/receipts capture

Day	Golden Glaze	FlexStay	AI / systems	Output
Day 1	Lock shop list, manager contacts, POS source per shop, and current cash-control process.	Lock active unit list, market, timezone, PMS ID, owner/vendor contacts.	Create one source-of-truth data map for both businesses.	data map v1
Day 2	Audit last 7 days of sales, cash logs, deposit bags, open shop issues, and existing investor/kitchen materials.	Audit today board, check-ins/check-outs, open issues, unread guest/owner/vendor threads.	Mark missing fields that block reliable daily rollups and raise materials.	missing-data punch list
Day 3	Define daily shop scorecard: sales, cash variance, issue count, check-in status, labor risk.	Define daily property scorecard: occupancy status, turnovers, P0/HIGH issues, message queue.	Standardize severity labels and escalation language.	daily scorecard template
Day 4	Fix cash variance denominator and map register/deposit mismatch by shop/date.	Confirm cleaning/vendor statuses are visible for same-day turnovers.	Update rollup logic to separate missing data from true exceptions.	variance + turnover report
Day 5	Review Golden Glaze high-priority issues and assign owner/date/status to each.	Review FlexStay P0/HIGH issues and vendor follow-up queue.	Create the "attention required" rules for both businesses.	attention rules v1
Day 6	Shadow-run Golden Glaze daily summary from real data; manually verify every number.	Shadow-run FlexStay daily summary; manually verify guest/vendor/turnover items.	Log wrong numbers, missing fields, and false alarms.	shadow-run error log
Day 7	Summarize week 1 misses: cash, inventory, issues, check-ins, payroll visibility, raise/kitchen data gaps.	Summarize week 1 misses: turnovers, messages, maintenance, vendor follow-up.	Pick the top 3 workflow fixes for week 2.	week 1 operator review
Day 8	Build or tune Golden Glaze missed-check-in watcher.	Tune FlexStay same-day turnover and vendor acceptance flags.	Add clear owner/escalation fields to every alert.	watchers v1
Day 9	Build cash variance review flow: POS expected, counted cash, bag count, deposit status.	Build maintenance triage flow: severity, property, vendor needed, next action.	Draft escalation prompts in Jack's voice.	two triage flows
Day 10	Pull payroll/labor source and identify why rows are missing.	Review guest response times and draft quality from last 20 usable messages.	Add "draft only / needs Jack YES" gate where money or policy is involved.	labor + draft gate notes
Day 11	Define inventory low-stock signals for top items that actually hurt revenue.	Define vendor follow-up cadence for cleaning, maintenance, and owner updates.	Connect recurring reminders to the right Discord channels.	inventory/vendor cadence
Day 12	Run first full Golden Glaze weekly packet: sales, cash, issues, labor visibility, risks, and raise/kitchen milestones.	Run first full FlexStay weekly packet: occupancy board, issues, messages, vendors, risks.	Compare packet to source systems and mark trust level by section.	weekly packet v1
Day 13	Fix any Golden Glaze packet section with low trust.	Fix any FlexStay packet section with low trust.	Turn fixes into checklist items so they do not recur.	trust fixes
Day 14	Quantify time saved from Golden Glaze summaries and exception routing.	Quantify time saved from FlexStay rollups and message triage.	Document one public-safe operator OS receipt.	week 2 review + receipt
Day 15	Promote reliable Golden Glaze alerts from shadow to supervised use.	Promote reliable FlexStay alerts from shadow to supervised use.	Define what the system may do automatically vs draft only.	supervised-use rules
Day 16	Build shop-level issue aging report: what is open, who owns it, how long.	Build property-level issue aging report: severity, vendor, guest impact, next touch.	Add stale-item escalation after 24/48/72 hours.	aging reports
Day 17	Review QBO class/location mapping for Golden Glaze finance reporting; define how raise proceeds and kitchen capex will be tracked.	Review QBO class/location mapping for FlexStay finance reporting.	List finance items that require bookkeeper review, not agent judgment.	finance mapping notes
Day 18	Create Golden Glaze manager-facing daily close checklist.	Create FlexStay vendor-facing closeout checklist for maintenance/cleaning tasks.	Make checklist completion visible in rollups.	closeout checklists
Day 19	Run side-by-side: manager manual report vs AI generated report.	Run side-by-side: manual property board vs AI generated rollup.	Measure difference and decide what can be trusted.	manual vs AI comparison
Day 20	Clean up false-positive Golden Glaze alerts.	Clean up false-positive FlexStay alerts.	Set alert threshold: only urgent, financial, guest-impacting, or aging exceptions.	alert threshold v2
Day 21	Quantify Golden Glaze weekly wins and remaining manual work.	Quantify FlexStay weekly wins and remaining manual work.	Choose next two workflows to automate.	week 3 review
Day 22	Start Golden Glaze inventory/reorder workflow for highest-impact SKUs.	Start FlexStay owner/case-manager follow-up workflow.	Keep both in draft/supervised mode.	new workflows started
Day 23	Review employee/overtime visibility and identify the fastest clean data path.	Review cleaner/vendor reliability by property.	Add trend view: repeated misses by shop/property/vendor.	trend report
Day 24	Build investor-safe Golden Glaze summary without hype: store base, $1M raise plan, central kitchen rationale, use-of-funds, and current risks.	Build owner/internal FlexStay summary with exceptions and next actions.	Separate public-safe, internal, and sensitive data views.	summary templates
Day 25	Run 30-day Golden Glaze scorecard draft.	Run 30-day FlexStay scorecard draft.	Reconcile with actual source systems before trusting.	30-day scorecard draft
Day 26	Fix top 3 Golden Glaze workflow gaps.	Fix top 3 FlexStay workflow gaps.	Document what still requires Jack vs manager vs vendor.	gap closure list
Day 27	Test if Golden Glaze can run a day with only exception-based Jack review.	Test if FlexStay can run a day with only exception-based Jack review.	Log every time Jack still had to manually hunt information.	exception-only test
Day 28	Calculate hours saved and misses caught for Golden Glaze.	Calculate hours saved and misses caught for FlexStay.	Prepare before/after operator OS proof.	proof packet
Day 29	Decide next Golden Glaze scale workflow for 7 → 10 shops and central kitchen execution.	Decide next FlexStay scale workflow for 11+ units.	Set month 2 priorities from measured pain, not guesses.	month 2 priority list
Day 30	Publish internal 30-day Golden Glaze AI ops review.	Publish internal 30-day FlexStay AI ops review.	Decide whether the plan is saving enough time to expand, pause, or refocus.	30-day decision memo

month 2

Promote reliable workflows to supervised operating rhythm. Expand Golden Glaze inventory/labor controls, raise packet support, central kitchen model, and FlexStay vendor/owner follow-up.

month 3

Build stronger weekly packets and trend reporting. Measure actual hours saved, errors avoided, and response-time improvements.

month 4

Use the system to support Golden Glaze 7-to-10 shop scaling, centralized kitchen rollout, and FlexStay multi-market operating consistency.

month 5-6

Standardize the operator OS playbook. Only then consider applying it to another acquisition, client, or external product.

scoreboard

metrics that matter

business outcomes

Golden Glaze: daily sales visibility, cash variance rate, register/deposit mismatch count, open high-priority issues, labor/overtime visibility, store check-in compliance.
Golden Glaze capital plan: $1M raise packet readiness, kitchen budget, equipment list, capex tracking, production capacity, transfer accuracy, waste, and launch milestones.
FlexStay: same-day turnover readiness, unresolved P0/HIGH issues, guest response time, vendor acceptance rate, vacant/in-house board accuracy, owner/case-manager follow-up queue.

AI system health

approval rate: percent of drafts Jack or managers can send with minimal edit.
error rate: wrong property, wrong shop, wrong business, bad number, or unauthorized commitment.
hours saved: recurring manual task minutes removed or reduced each week.
escalation quality: urgent items surfaced quickly without flooding Jack.

hard rules

guardrails

never send FlexStay guest replies without Jack's approval or explicit workflow permission.
never auto-reply to Golden Glaze customer DMs.
never quote prices, comps, refunds, exceptions, or lead times without Jack.
never split Golden Glaze and Donut into separate businesses.
never treat the $1M raise or centralized kitchen as a separate business; they are Golden Glaze growth initiatives.
never present Golden Ops or selling premium as operating businesses in this plan.

positioning

how to talk about it

Jack is building an AI-enabled operating company, not an AI agency first. The public story is: "I run a 7-shop donut rollup and 11-unit housing business, and I am using AI agents to replace the messy SaaS stack and reduce my direct operating load." The Golden Glaze growth story includes a planned $1M raise and centralized kitchen, supported by operating proof instead of hype. The proof is the operating footage, dashboards, rollups, and before/after time savings.

next 7 days

Golden Glaze

Finish cash variance denominator, register/deposit mismatch mapping, payroll visibility, and the first raise/kitchen punch list. Keep open issues tied to shop, severity, owner, and date.

FlexStay

Keep the 8am/6pm rollups useful: same-day turnovers, vendor cleaning status, P0/HIGH issues, unread owner/case-manager/vendor messages.

Content

Publish the operator OS story through real receipts: store ops, housing ops, dashboards, mistakes caught, hours saved. No generic AI agency positioning.