Charleston, in person, two weeks · supersedes the 2026-05-20 research report · planning for Mark only (not for sharing with William)
William invited you to spend two solid weeks in Charleston at the property — you stay in the apartment above the shop, Hannah lives in the renovated house adjacent, William is building on Monk's Island. That changes the entire calculus. The original "2-4 remote weekends" plan was always going to be slow and bandwidth-limited. Two weeks of in-person, side-by-side work with Hannah, with a fully prepped data set on arrival, is the version of Option A that actually ends with Priority running on LMN.
The constraint is no longer time. It is how much you arrive prepared with. Two weeks of fumbling around in LMN with Hannah ends okay. Two weeks where days 1-3 are loading pre-built Priority-specific budget, service catalog, and customer list ends with a live system. The lever is the 4-5 weeks of remote prep, done one Bidet questionnaire at a time, so William never gets overwhelmed.
The Hannah-overwhelm guard from the first plan stands. One questionnaire per week. After Charleston, one weekly 30-min check-in for a month, then we re-decide. If at any point you or Hannah says "this is too much," we stop and re-scope. Friendship over project.
The goal, in one sentence
By the end of the second week in Charleston, all six Priority crews are clocking in through LMN, Cristen is invoicing through LMN, Hannah owns the office workflow, and you fly home with a 1-page operations checklist she runs every week without you.
Phase 1 — The prep, 4-5 weeks remote, one questionnaire at a time
Send William one Bidet brain-dump prompt per week. He records whenever, however scattered. Bidet cleans it; you read it; I organize it into LMN-import-ready data. By the time you fly to Charleston, the heavy data-entry pain that usually eats LMN's first month is already done in a staging account.
The six questionnaires, in order of low-effort → high-impact. Each is a 15-25 minute brain dump for William. The exact prompts below are copy-paste-ready — text them to him one per week:
Week 1Service catalog15 min · easy · William loves this
"Walk me through the 10-15 job types you actually do, the ones that come up over and over. Big paver install. Maintenance route. Irrigation install. Sod replacement. Whatever the regulars are. For each one: roughly how long does it take, what crews handle it, what's the most common gotcha that costs you money. Just talk, no order, no structure."
Why first: William loves talking about the work, this gets him into the rhythm of using Bidet without it feeling like homework. The output becomes LMN's job templates — saves Hannah hours of typing later.
Week 2Crew composition + specialization15 min · easy
"Six crews. Talk me through who's on each one. Who's the foreman, who's the equipment guy, who's the laborer. What does each crew specialize in — install vs maintenance vs hardscape vs irrigation. Which guys are best at what. Who needs supervision, who can be left alone. And the truck assignments — which crew drives which rig."
Becomes LMN's crew assignments + the routing logic. Also tells me which crew should be the pilot when we go live (probably Connor's, since he's a Project Manager and gives clean feedback).
Week 3Customers — the top 30 + the painful ones20 min · medium
"Top 30 customers by revenue, ballpark. Residential vs commercial mix. Which are recurring (maintenance route) vs one-shots (install). For the top 10, what makes them good customers — do they pay on time, do they refer, do they let you upsell. Then the painful ones: who's slow to pay, who's a chronic scope-changer, who do we keep because of the address even though the work is hell. No need for names if it feels weird."
This shapes LMN's customer segmentation and helps Cristen's CS module. The painful-customer signal is gold for what to flag in LMN's communication thresholds.
Week 4Cost structure — the LMN budget module's whole point30 min · harder · needs Cindy + Hannah
"Your true cost per man-hour, with everything: wages, payroll taxes, workers comp, health insurance, vehicle costs (purchase, fuel, maintenance, depreciation), equipment costs, fuel, dump fees, materials markup, office overhead, insurance, software, the works. If Cindy's QuickBooks has a P&L for the last 12 months, that's the gold mine. Hannah can help assemble. The goal: one defensible number for cost-per-billable-man-hour. LMN's whole estimate engine pivots on this."
This is the one I'd schedule with Hannah specifically. William can talk through the framing in 10 minutes, then Hannah + Cindy actually produce the number from QuickBooks. If we have this dialed in by week 4, the entire Charleston trip's data work is anchored.
Week 5Pricing logic15 min · easy
"Walk me through how you actually price. Material markup percentage. Labor multipliers. Design fee structure. Change-order policy. Do you discount for recurring customers, for cash up front, for referrals. When do you walk away from a job because the customer's price expectation is wrong. Just the rules in your head."
Becomes LMN's estimate templates. Combined with Week 4's cost structure, every quote LMN produces is anchored to real Priority margins.
Week 6 (optional, if time)Three good margins, three almost-losses20 min · storytelling
"Three jobs in the last year with great margins — what made them work. Three that almost lost money — what went wrong. Just stories, not a P&L. Where did the estimate miss the reality. Where did scope creep. Where did a vendor screw you. The pattern in these stories is what LMN's job-costing module is supposed to catch in real time going forward."
Optional but valuable. Becomes the training material for what to FLAG inside LMN going forward — "job exceeds estimated hours by X percent, alert Connor." Skip if the week is busy.
What you're doing during prep, beyond reading dumps: getting your own LMN demo account (free trial, ask Granum) and rehearsing the office workflow. Configuring the staging account with each questionnaire's output as it comes in. By the time you fly, you've personally clicked through the budget module a hundred times and know its quirks cold. Hannah's first contact with LMN is with you sitting next to her at her desk, not alone in front of a video.
Phase 2 — Two weeks in Charleston, day by day
Assumes you arrive on a Saturday. Adjust day labels if you fly mid-week. The shape is: setup & relationships → office team → pilot crew → rollout → handoff.
Week 1 — office + pilot
Day 1 · SatArrive, settle in. No software.
Property tour. Apartment keys. Coffee with William and Cristen, dinner if they're up for it. Stop by Hannah's house on the property to say hello. The whole day is relationships, no laptops. The trip succeeds because they trust the person, not the software.
Day 2 · SunWalk the property. Map the operational reality.
Look at the trucks, the shop, the equipment yard. Spend time with William at the new house build on Monk's Island if he's heading out. Casual conversation only — you're filling in the picture, not interviewing.
Day 3 · MonFirst office day. Hannah audit.
Sit with Hannah at her desk. What does she use today? Email setup, calendar, customer files, whatever billing tool, paper notes. Two hours of "show me how you do Tuesday." Take notes. Cristen joins for the customer-service slice. No LMN today. The goal is to know what you're replacing, not to start replacing.
Day 4 · TueLoad the prepped data into Hannah's live LMN. Office training day 1.
Customer list import (300 properties, prepped). Service catalog (loaded from Week 1 questionnaire). Budget module set up with Week 4 numbers. Walk Hannah through navigating the dashboard, finding a customer, finding a job. Light touch — she's seeing, not doing yet.
Day 5 · WedOffice training day 2. Cristen on CS module.
Cristen takes the customer-service module — communication threads, the customer portal setup. Her 9 years at Verizon make this fast. Hannah practices producing one real estimate from a real call. End of day: she's produced an estimate that William looks at and goes "yeah, that's right."
Day 6 · ThuCrew app demo with Connor. Pilot setup.
Spend a couple hours with Connor in the yard. Show him the crew app on his phone. Walk through clocking in, taking job photos, marking complete. He's your pilot. Get his honest read on UX friction. One critical question: does the cell signal at job sites hold up? If a site is in a dead zone, the crew app needs to offline-queue. Verify before live rollout.
Day 7 · FriLight day. Live triage and prep for week 2.
Cristen has now sent a few customer comms through LMN. Hannah has produced 2-3 estimates. Triage the rough edges. Set up Joey's maintenance crew for next-week onboarding. Dinner with William and Cristen if they want — you've earned it.
Week 2 — rollout to all six crews + handoff
Day 8 · SatOff, if you can.
A real day off. Beach. Sleep. The next 6 days are dense.
Day 9 · MonConnor's crew goes LIVE. Other foremen observe.
Connor's crew clocks in on the app for real, runs their day in LMN. You + Hannah are on-call. Other foremen (Joey + the other 4 install/maintenance crew leads) shadow Connor at end of day, see how it went.
Day 10 · TueJoey's maintenance crew onboards.
Joey is the senior guy and his maintenance route is parallel work to Connor's install — good test of LMN handling two job-types simultaneously. End of day: 2 crews live, 4 remaining.
Day 11 · WedCrews 3 + 4 onboard.
Two crews in one day if both foremen pick it up cleanly. Otherwise drop to one.
Day 12 · ThuCrews 5 + 6 onboard. All six live.
By end of Thursday, all six crews are clocking through LMN. Hannah is producing estimates. Cristen is sending invoices. The system is running.
Day 13 · FriTriage day. The 'what's broken' list.
Spend the whole day fielding "this isn't quite right" from every role. Fix what you can. Document what needs LMN support. Build a 1-pager: Hannah's weekly checklist — the 5-7 things she runs every Monday to keep the system clean. This is the handoff artifact.
Day 14 · SatWrap. Fly home.
Coffee with Hannah to walk through the weekly checklist. Hug William and Cristen. Get out before you become permanent IT support. The handoff is real or it isn't — this is the day it gets tested.
Phase 3 — Back home, weeks 3-8
Week 3-6: Weekly 30-min Hannah video call. What's working, what isn't. Triage as needed. Most issues at this point are LMN support tickets, not Mark tickets.
Week 4-5: Ship the customer-comms autopilot as the first Option B augment. Cleanest match to Priority's one real review complaint (communication). Uses the framework already built for Bidet.
Week 7-8: Reassess. If everything is humming, decide whether the voice-to-estimate and photo-to-job-scope augments are worth shipping or if LMN out-of-the-box is enough.
Month 3+: You're an external advisor at most. Hannah owns it.
Hannah's hardware
Add this as a quick early-phase question to William: what does Hannah use today? If she's on a 5-year-old laptop or worse, this trip is a great excuse to upgrade. LMN runs in browser plus their crew mobile app — doesn't need a powerhouse, but it needs to be modern, reliable, and not slow her down all day. Budget $1,500-2,500.
If she's Mac-curiousMacBook Air M4, 13", 16GB RAM, 512GB SSD — ~$1,400. Silent, all-day battery, runs everything she needs forever. Cristen + William's other tech might already be Apple; check.
If she's Windows-committedDell XPS 13 (Snapdragon X Elite) ~$1,300 or ThinkPad X1 Carbon Gen 13 ~$1,800. Both are quiet, durable, fast enough.
External monitorWhatever she has + a 27" 4K (LG/Dell ~$300). LMN is dense; she'll want screen real estate.
SkipGaming laptops, anything with a discrete GPU, anything called "Workstation." Overkill, expensive, hot.
Order it BEFORE you fly to Charleston so it's set up Day 3 instead of waiting on shipping mid-trip. If William or Hannah have a vendor preference, let them order; if not, B&H (which we just used for your Apex parts) ships fast and clean.
Claude plan — who needs what
Mark (you)Claude Max / your current setup. No change.
HannahIf she wants to use Claude for her own daily work (drafting emails to customers, summarizing meeting notes, writing reports for William): Claude Pro at $20/mo is the right tier. She'd use Claude.ai web. No team license needed unless...
If you want shared ProjectsClaude Team, $25/user/month, min 5 seats = $125/mo. Gets shared "Projects" with custom instructions, shared chat history, central billing. Useful if Hannah + Cristen + you all want to use Claude on Priority's stuff with shared context. Probably overkill at this scale — revisit if it's working in 6 months.
Claude EnterpriseCustom pricing, SSO, SOC 2, audit logs. Not a fit until Priority is 50+ employees or has compliance needs. Skip.
Anthropic APIFor the customer-comms autopilot I'd build for them — that's API access, billed pay-as-you-go. Likely under $30/mo at Priority's volume. Could be billed to your existing API key or to William directly (recommend William directly for cleanness).
Net recommendation: Claude Pro for Hannah ($20/mo), possibly Cristen too. Hold off on Team. API gets its own key, William's billing.
Antigravity — the more honest read
You called me out for over-dismissing Antigravity, fair. The accurate read: Antigravity 2.0 has real multi-agent capability (Mission Control, browser agent, parallel-task orchestration) that ChatGPT also has now — agentic dev is becoming table stakes. Where Antigravity actually shines for your case:
Parallel-research bursts. Want to evaluate 4 different LMN-import approaches at once? Spin up 4 Antigravity agents in parallel, pick the best. Where Claude Code is serial-by-default, Antigravity is parallel-native.
Greenfield prototyping. The customer-comms autopilot prototype (Option B, ship in Phase 3) is a clean sandbox — not yet on William's prod, no real-customer downside. Good Antigravity use case.
Browser-agent demos. Its browser-agent is genuinely cool for "show William what LMN setup will look like" walk-throughs.
Where I still hold the original line: not for the customer-facing production work on William's livelihood, this quarter. The reported instability (suspended accounts, lost connections, the local-drive-deletion incident) means anything load-bearing for Priority stays on Claude Code until Antigravity has a GA stamp. Keep it installed. Use it on the prototype. Revisit promotion for prod in Q3 if it stabilizes.
The Hannah-overwhelm guard, restated
This is the part that can blow up a friendship if I don't say it clearly: if at any point William, Cristen, or Hannah feel "this is too much, too fast," we stop. The two-week timeline is a north star, not a contract. The questionnaires are spaced one per week BECAUSE William's processing bandwidth is finite, and his Bidet-shaped brain (your words) won't push back when he's overloaded — he'll just go quiet. If a week goes by without a questionnaire response, that's the signal to pause and check in. Don't send the next one until he's caught up.
Same in Charleston. If Day 4's data load takes two days instead of one because Hannah needs more breathing room, we shift. The plan is a story, not a Gantt chart.
What this teaches us about Legacy Soil — updated
The "Hannah" role is the lever. Priority's success depends on a daughter who owns the system day-to-day. For Legacy, that one-ops-person role doesn't exist yet — right now it's all you. As Legacy scales past the founder, identifying or hiring "Legacy's Hannah" is the question. Until then, every system you choose has to be one you can run alone, including weeks when you're at school until 11pm.
Brain-dump questionnaires are a transferable pattern. The six William prompts above are basically the structured-interview-by-async-voice approach. When Legacy eventually has its first real customer pipeline, the same approach works for capturing their stories before composting completes — "tell me about [pet], take your time."
In-person time has compounding ROI. A two-week sprint in person beats six months of weekend remote. For Legacy: the in-person customer experience — the family visit, the pearl ceremony — is what compounds. Spend the build cycles on those, not on the ops core.