01 · Thesis

We build the technology layer your business would never build for itself

Our clients either earn enough per hour that a $15–25k buildout is obviously rational (the Chief of Staff line), run operations at a scale where a custom platform pays for itself in months (the Ops line), or need private data and intelligence their industry can't buy off the shelf (the Data & Intelligence line). One firm, three lines, same delivery team — different vocabulary and sales motion per line.

Every YouTube tutorial that says "set up Claude in an hour" reinforces our market. The setup friction is not a bug we need to solve — it is the reason the service exists. Chief of Staff clients pay us because 12–20 hours of concentrated setup across multiple sessions is time they cannot spend. Ops clients pay us because building the platform that actually runs their floor is a nine-month project their team was never going to start. Data & Intelligence clients pay us because we already maintain the scrapers, pipelines, and datasets their in-house team would take years to build.

Service lines

Same firm, same delivery team, same tech stack. Different vocabulary, different buyer profile, different sales motion per line. New verticals (Healthcare ops, Legal ops, etc.) fold under one of the three existing lines — or get their own line once a real case closes. Not before.

02 · Principals

03 · Target customer

Wealthy, well-connected, non-technical people whose time is worth far more than the service costs them. Their gap isn't a tool — it's a trusted person who handles the AI layer of their life.

Who they are

• Net worth $5M+ (with real stretch capacity past $100M)
• Business owners, partners at professional firms, family-office heads, successful operators
• Already pay $200–500/hr for their lawyer and accountant without thinking about it
• Chief of Staff at this level costs $167–400k/yr salary; we sit alongside that budget line
• Value discretion and control more than cost savings

Who they are not

• Anyone shopping a price page
• SMB owners looking for a Zapier freelancer
• Enterprise IT buyers — that's McKinsey or Slalom, not us

04 · Positioning

Premium AI boutiques (AE Studio, Artefact, QuantumBlack, Palantir) publish no pricing, no public tiers, no form submissions. Only down-market firms do. We copy the premium pattern.

The phrase

"AI Chief of Staff" — we use a role name the clientele already pays for (typically $167–400k/yr for a human chief of staff) as the reference anchor. The word "AI" appears later in the conversation, if at all.

Language rules

Sales principles

• Doctor, not pharmacist. We diagnose first, prescribe second. Sometimes the right answer is a $30 SaaS we recommend without selling a build. Honesty in the wealthy-network referral chain compounds; over-selling poisons it.
• Minimum AI possible. When a problem is solved by a non-AI automation or an existing tool, we say so. Reinforces that we're an outcome firm, not a tech-reseller.
• "You don't have a lead problem — you have a follow-up problem." Punchy reframe for any business owner complaining about pipeline or ad spend. Sets up the Voice Agent conversation.
• Strategic advisory beats tool reselling. The margin is in picking the right configuration for the client, not in being the tool. Confirmed across every serious firm we've studied.

Marketing posture

• No public price page. Ever. Pricing lives in private discovery.
• Case studies name outcomes, never dollar figures. "Reclaimed 12 hours a week," not "$75k saved."
• "We take on 4 new clients per quarter." Scarcity signal that's currently accurate.
• Intake is referral-gated. No form submissions, no "book a demo."
• Tier names are internal shorthand. Justine says "you'd be a Lambo client" at dinner; we never write it down for a prospect.

05 · Engagements

We are a project firm with a redeployable parts library and a small junior-labor delivery team. Most engagements are bounded projects with a support tail; retainers exist as an option for clients who specifically want one. All names below are internal shorthand — the prospect sees a custom scope, not a menu.

The product line

The Test Drive specification

Family / Estate Operating System — the upsell after a first win

This is not a gateway product. The path is: client does a first engagement (Quick Setup, AI Audit, Test Drive, or single project) → feels one win → Justine or Curtis raises Family/Estate OS as the natural next step → scoped at $25–35k for 5–7 weeks of work.

Includes a pre-built skill set covering travel/concierge research, household-staff coordination, philanthropy and charity admin, unified family calendar across 3–4 separate calendars, voice-matched personal correspondence, and receipt/expense capture. 60-day post-build support window included. After that, growth happens through $2–5k add-ons, not a monthly retainer — Family/Estate clients have just paid $25–35k and won't double-pay for ongoing service.

Why $25–35k is the right band: wealthy clients are more allergic to feeling like they got the starter version than to a high price tag. Replaces or augments a personal assistant ($80–150k/yr salary). Same psychology as the existing "chief of staff at $167–400k/yr" anchor, operationalized for family/personal life rather than business life.

$35k applies when multiple businesses interleave with family life, MFA-heavy estate stack with complex authentication, second principal involved (spouse with their own correspondence/calendar), or on-prem privacy requirements (local AI box typically separate).

Retainers — for clients who specifically want one

Some clients do want a retainer — workaholic operators, family offices that want a true on-call team, business owners who treat us as a permanent line item. For them, three retainer tiers stay on the books. This is not the modal engagement — we don't lead with retainer in proposals. Path into retainer: prospect specifically asks for one, OR post-build add-on volume justifies consolidating.

Framing rule: a retainer is continuous access, not a subscription. Build-heavy months run hot, tune-and-respond months run light, the monthly fee smooths it. If a client starts counting hours, we've positioned this wrong.

Support tails after any major build

Wealthy clients don't want to keep paying once the work feels done. A retainer feels like a tax; a support tail with a clear endpoint feels like service. Post-build motion for Test Drive / Family-Estate / Voice Agent / custom projects:

• 30-day soak — included in any build over $5k. Bug fixes, broken connectors, small tweaks.
• 60-day support — included in Family/Estate OS price.
• 90-day extended care — optional, $1,500–2,500 flat.
• Per-incident — $250–400 per fix after all soak windows close.
• Add-on builds — $2,000–5,000 each, scoped à la carte. New skill, new automation, new connector.

Most Family/Estate clients add 2–4 things per year at $2–5k each — that's $4–20k/yr in expansion without the client ever feeling overcharged. This is the right shape for our buyer.

06 · Pricing & the Book of Redeployable Parts

Pricing only makes sense paired with the model behind it: a small partner-led firm with a junior-labor delivery layer, leveraging a shared library that compounds with every engagement. The Book of Redeployable Parts is the actual moat — and the reason moderate prices produce strong margin.

The Book of Redeployable Parts

Every productized engagement contributes a skill, a connector pattern, a prompt template, an install script, or a context map to a shared internal library. Build #1 is full price + full hours. Build #2 onward reuses the library:

Margin compounds. Prices stay anchored. We never lower a price because we got faster — faster delivery is our margin, earned from doing the work to standardize it. Clients pay for the standardized result, not the time it took.

Labor model — two rails plus the library

• Curtis (partner labor) — discovery, scoping, complex builds, handoff, anything that touches a wealthy principal directly. Internal rate floor: $500/hr. Never quoted externally.
• Junior delivery (NAIT hires, 4–6 month rolling) — standardized work pulled from the parts library, build execution, testing, install support. Wage $35–45/hr (~$50–55/hr fully loaded). First hires happen after the first 2–3 paid client engagements — Curtis runs them solo to prove the parts library, then scales delivery capacity.
• The library — every productized engagement adds to it; every subsequent engagement of the same shape benefits.

Margin math at build #2+ (the steady-state shape)

AI Audit margin is intentionally tight — it's a paid filtering tool. Tells us whether the prospect is real, gives us deep context for the next engagement, pre-sells at full margin. Most audits convert into a Test Drive or larger build within 60 days at full margin (audit fee credits in full).

Quick Setup margin is intentionally moderate — Curtis can't fully step out of any client interaction even on small jobs. The chain math (below) shows why we run it anyway.

The chain math — why cheap entry products are worth running

A single Quick Setup that converts is worth far more than its $1,500 sticker. Worked example, conservative assumptions:

The margin per Quick Setup is small. The chain is the business. This is why we run a $1,500 product at all — and why every productized engagement has a clear converted-next-step that builds on the parts library.

Hardware — markup + setup time

We source, prep, configure, and deploy AI compute for clients who want local inference or on-prem privacy. Hardware cost + 30–50% markup (our sourcing margin) + setup time billed per scope. Typical deployments: small AI box on the desk (Mac mini or equivalent), workstation with GPU for local LLMs / vector databases, or whatever the stack needs.

What we don't install ourselves (but recommend trusted vendors for): camera systems, network wiring, structured cabling, smart home integration, ISP/tenant admin, printer fleets.

The pricing-rationale conversation — for clients who push back

• Compared to a personal assistant ($80–150k/yr): the same outcome at a fraction of the cost, with no hiring risk or turnover. The most common pricing anchor for our wealthy buyer.
• Compared to a chief of staff ($167–400k/yr): same anchor, business-side. Use this for SMB-owners.
• Cash flow / risk reframe: "Revenue minus costs is the obvious lens. The underrated lens is cash flow over risk. Some AI tools increase cash flow but also increase risk — we focus on the ones that reduce risk in your operations while improving throughput. That's what makes a system worth paying for." Particularly effective with family-office and risk-averse buyers.
• 10x ROI in 12 months: internal sanity check. If a build can't plausibly return 10× its price within a year (in time saved, revenue captured, or risk reduced), scope down or change the project.

07 · Engagement flow

Referral in. Private discovery. Scoped written proposal. First engagement. Support tail. Add-ons over time. Retainer only if the client specifically asks for it.

1. Referral or introduction. Justine or Josh mentions us in conversation, describes the outcome (not the tool), lets curiosity build, asks if the prospect wants a private intro.
2. 30-min intro call with Curtis. No pitch — just listening. We're deciding whether there's a fit. They're deciding whether we're serious.
3. Discovery conversation (45 min). Run the intake — understand their stack and pain surface. Includes the anti-to-do list exercise (every task they never want to manually do again becomes a candidate skill) and the progressive trust conversation (Drafts → Read-only → Confirm-to-act → Autonomous; most clients live in stages 1–2 for the first six months). Sales AND interviewing at the same time.
4. Written proposal. One scoped engagement (Quick Setup, AI Audit, Test Drive, Family/Estate OS, Voice Agent, or custom build) + fixed price, delivered in one page within 24h of the discovery.
5. First engagement. Build, install, train, soak. Client is using the deliverable in daily use before we step back.
6. Decision point. Most clients enter the support-tail motion: 30/60-day soak included, optional 90-day extended care, then per-incident or add-on builds as needs emerge. Retainer only if they specifically ask. Some clients choose clean exit and become referral sources — they keep the automation either way.
7. Layered-in growth. New skills, automations, connectors, hardware installs, strategy work — quoted as scoped add-ons ($2–5k each). For clients who've done a basic engagement, Family/Estate OS is the natural next step ($25–35k for 5–7 weeks).
8. Retainer (optional path). If the client's add-on volume justifies it, or they specifically want one, consolidate into Daily / Weekend / Lambo. The AE (once we have them) or Curtis takes ownership.

Fee-credit-on-conversion

Borrowed from Concierge Studio, the firm structurally closest to ours. Test Drive (and Quick Setup, and AI Audit) fees apply 100% against the next engagement if signed within 60 days. De-risks the entry without cheapening the price.

08 · Growth loop

Curtis solo for the first 2–3 paid clients to prove the parts library on real engagements. Then hire NAIT juniors on rolling 4–6 month placements as a delivery layer. Curtis stays on discovery, scoping, and any work that touches a wealthy principal directly. The library compounds; junior delivery scales.

Phase 1 — Curtis solo (first 2–3 paid clients)

• Curtis runs every engagement end-to-end. The point isn't capacity — it's building the parts library against real clients.
• Every productized engagement (Quick Setup, Test Drive, Voice Agent, Content-Ingest, Family/Estate OS) ships a reusable skill, connector pattern, prompt template, or install script back to the library.
• By client #3, build #2 of any given product type is already library-assisted. Hours start dropping.
• First-year revenue per client (mixed mix of Quick Setup, Test Drive, occasional Family/Estate OS, plus add-ons): expected $15–40k/client blended.

Phase 2 — first NAIT hire (after the first 2–3 clients land)

• Hire trigger: parts library proven on 2–3 clients AND Curtis is at his solo time-ceiling.
• NAIT junior on a 4–6 month placement, 20 hrs/week to start, scaling toward 40. Wage $35–45/hr (~$50–55/hr fully loaded).
• Junior owns standardized productized delivery: Quick Setup builds from SOP, Test Drive workflow execution, Voice Agent configuration, Content-Ingest skills, Family/Estate OS skill-building under Curtis's scoping.
• Curtis stays on: discovery, scoping, voice tuning, install, anything client-facing on a wealthy principal, audit deliverables, complex custom builds.

Funding math (project-driven, not retainer-driven)

• One Family/Estate OS at $25k library-assisted = ~$18–20k contribution margin (build #2+).
• One Test Drive at $15k library-assisted = ~$9–11k contribution margin.
• One NAIT junior at 20 hrs/wk = ~$1,000–1,200/wk loaded cost = ~$4,500–5,500/mo.
• One Family/Estate OS funds ~3–4 months of one junior's capacity. Even one Test Drive funds ~2 months. Junior layer is self-funding from the second productized engagement onward.
• Add-on revenue layered on top: most served clients add 2–4 things/yr at $2–5k each, with junior delivery handling most of the build.

Capacity ceilings

• Curtis solo: ~8–12 productized engagements/yr before burnout (more if mostly Quick Setup; fewer if mostly Family/Estate OS).
• Curtis + 1 junior at 30 hrs/wk: capacity roughly doubles for productized work, leaving Curtis the partner-level conversations.
• Curtis + 2 juniors: ~25–35 productized engagements/yr plus a steady stream of add-on builds.
• Family/Estate OS and any custom build with a wealthy principal always require Curtis directly; juniors assist on the build side but don't own client relationship.
• Retainer clients (when they happen) are an exception: Curtis owns Lambo, AEs (post-junior, eventually) take Daily/Weekend.

Land-and-expand within wealthy-network accounts

The other half of growth — beside new logos — is expansion within a single principal's orbit. Wealthy clients live in webs: family office, operating company, foundation, household staff, spouse's separate businesses. A first engagement that lands well is rarely the last engagement; it's the first foothold.

The expansion pattern, named so we deliberately look for it:

1. First foothold. Quick Setup, AI Audit, or Test Drive on the principal directly. Lands one win, builds trust.
2. Adjacent role. Spouse, EA, Chief of Staff, or board secretary asks to be set up next. Often surfaced organically as the principal talks about us at home or at work.
3. Adjacent function. Foundation administrator, family-office controller, or operating-company ops lead — same firm, different department, different problem set, often a custom build or Family/Estate OS scope.
4. Adjacent entity. Principal mentions us to their business partner, a friend at another family office, a board peer. New first foothold, same chain plays again.

This is not cold expansion — every step is referral-warm because the principal vouches. The growth ceiling per relationship is high: a single $1,500 Quick Setup landing well at the right family office can compound into $50–150k of work over 2–3 years across spouse, foundation, ops, and board correspondence. Justine and Josh should explicitly listen for adjacent-role and adjacent-entity openings during regular conversation with current clients. The opening usually appears casually and gets missed without a deliberate ear for it.

09 · Talent strategy

Our budget ($30–40/hr CAD, $62–83k FT) doesn't land CS PhDs or FAANG engineers — and we don't need them. Research shows the best-fit archetype is not what most consulting firms hire.

Who we hire

Required skills from day one

• Python (Flask, APScheduler — our stack)
• OAuth + Google Workspace APIs
• Claude API + MCP server basics
• SQLite / basic data modeling
• Non-negotiable: client-facing composure with UHNW clients. Technical gaps we train in 3 months. Social skills we can't.

First-hire triggers and structure

• Hire trigger: $15–20k MRR (first 3 clients retained)
• Base: $72k/yr CAD (~$35/hr at 2,000h)
• Bonus: 10% of retained-client revenue they directly service, paid quarterly
• W2 employee (not contractor) — trust signal for top-tier clients
• Benefits: basic health + dental
• Equity: none initially; phantom equity / profit share possible Year 2+

10 · Competitor map

A reference for the three of us and future employees. Full detail in the research archive; summary here.

The pattern we're copying

1. Named tier ladders are a down-market tell → tiers stay internal only
2. Entry engagement is paid, fixed-price, 2–3 weeks → our Test Drive matches
3. Project is the unit of sale; retainer emerges retroactively → don't lead with retainer
4. Upsells called "layered in" → adopted in language rules
5. Case studies publish outcomes, never dollar figures → adopted
6. Intake is referral-gated or discovery-call-gated → adopted

11 · Execution

Done

• Rate card v4 — productized engagements with margin math (internal, not for prospects)
• Full business plan with Book of Redeployable Parts as load-bearing concept
• Four research files: Gmail multi-account options, pricing benchmarks, boutique packaging, firm valuation + hiring
• YouTube-derived patterns synthesized into the plan (`research/youtube-lessons.md`)
• Path B spec + first implementation plan drafted
• This plan site live at hq.curtishall.ca
• Passcodes rotated fresh for the three founders
• Justine's personal Cowork + Gmail/Calendar setup live (case study #1 in progress)
• Numbered Alberta company incorporated; ATB business account opened

Active

• Justine in active discovery conversations with first wave of prospects
• Test Drive runbook + delivery SOP fully written (`hq-site/docs/test-drive-runbook.md`, `test-drive-delivery.md`)
• Anti-to-do list intake artifact + progressive trust framing baked into discovery script

Milestones

• M0: Plan v3 shipped. Productized engagements live. Incorp + banking complete.
• M1: First two paying clients delivered (Curtis solo). Parts library seeded with first reusable skills.
• M2: Third client delivered. Library is now reusable across at least 2 productized engagement types. NAIT junior hire trigger — Curtis at solo time-ceiling.
• M3: First NAIT junior onboarded (20 hrs/wk). Junior shadowing Curtis on first builds, then taking over standardized delivery from SOP.
• M4: Junior at 30–40 hrs/wk. Curtis back to mostly discovery + scoping + complex builds. First Family/Estate OS engagement landed (build #2 of the engagement type, library-assisted).
• M5: Steady-state on first cohort: 6–10 active client relationships in support-tail mode, regular add-on flow, second junior hire triggered.
• M6: 12+ clients across mixed engagement types. AE/FDE-track promotion conversation for the first junior. Acquisition-readiness checks against the exit-value criteria.

12 · Open questions

Decisions for the three of us. Drop comments under this section — or under any specific section — as you think through them.

1. Firm brand name. Currently unnamed. Options to ideate: Lantern, Atlas, Meridian, Signal, Concierge [X], something Curtis/Justine/Josh land on. Needed before first written proposal goes out.
2. Final ownership structure at incorporation. Under discussion on the Shareholders tab — see the current draft of the three-pots model + equity split options there.
3. Liability insurance level. What coverage do we need before touching UHNW inboxes? Get legal input before client #1 signs.
4. Contract templates. Test Drive agreement, retainer agreement, NDA — all TBD. Priority before the first signed engagement.
5. Referral credit for existing clients. If a Weekend client refers a Lambo client, do they get a credit? Industry-standard: 1 month free for successful intro.
6. Case study publication permissions. When and how do we get sign-off from Justine and first paying client to publish anonymized wins?
7. Marketing posture. Bare-minimum referral-only site vs. small content presence (case studies without dollar figures). UHNW clients value discretion; "barely a site" may be more prestigious than a built-out one.
8. Exclusivity of Justine and Josh. Is this their priority, or parallel to other sales / consulting commitments? Affects how we set expectations on throughput.
9. Justine's first month. Free pilot as client zero, or standard pricing with a discount? Recommendation: free 30-day pilot for her only (client zero investment), standard for every referred friend.