00 — START HERE: TPC Advisory Onboarding Kit v2 (Scoped Overlay)
Scoped overlay on kit-advisory-onboarding-v2. Holds TPC-cohort-specific data and defines the three operating modes members receive. For methodology, extraction protocol, output intelligence, and quality framework — go to the universal kit.
What This Is
The TPC-specific layer of the Advisory Onboarding Kit v2. The universal kit defines the methodology. This kit defines what TPC members actually receive and holds their practice-specific data.
Every member receives a kit with three operating modes. The kit produces client-facing onboarding deliverables (roadmap, action plan, mutual responsibility doc). The operating mode determines how much AI does.
The Three Operating Modes
Level 1 — Structure (The Member Runs It)
What the member gets:
- Documented onboarding SOP for their practice — every step, who does what, when
- Roadmap template — populated with their practice data, per-client fields left to fill
- Action plan template — same
- Mutual responsibility template — same
How it works: Member follows the SOP. For each new client, they fill in the per-client fields (name, start date, specific goals), produce the deliverables manually, and hand them to the client.
Why it matters beyond "the manual version": This IS the foundation. The SOP documents the process that AI needs for Levels 2 and 3. Without Level 1, there's nothing to automate. It also protects the practice — if AI is unavailable, the process still works.
Level 2 — Streamline (The Member Manages, AI Produces)
What the member gets: Everything from Level 1, plus:
- AI production prompts — specific instructions the member gives Claude to produce each deliverable
- Practice context file — a document Claude reads to understand this practice (team, tools, cadence, packages)
How it works: New client signs. Member opens Claude, loads their practice context file, and says: "New client: [Name], starting [Date], [Package Name]." Claude drafts the roadmap, action plan, and mutual responsibility doc. Member reviews, adjusts, sends.
What the member manages: Triggering each step. Feeding Claude the client-specific details. Reviewing every output before it goes to the client. Deciding when something needs their personal touch vs. when the draft is good enough.
Level 3 — Delegate (AI Runs It, Member Oversees)
What the member gets: Everything from Levels 1 and 2, plus:
- A kit instruction file — a full Claude instruction set (like a CLAUDE.md) that the member's Claude instance loads. Claude knows the practice, the team, the tools, the packages, the cadence, the welcome experience, the completion gate. All of it.
How it works: New client signs. Member says: "Onboard [Name], [Package], starting [Date]." Claude produces all three deliverables, flags anything it can't resolve (e.g., the first-win needs member input because it's judgment-based), and presents the package for review. Member reviews, approves, ships.
What the member oversees: The output. Does the roadmap reflect this specific client? Is the first win the right one? Does the mutual responsibility doc match what was discussed? The member is the quality check, not the production engine.
What Gets Produced Per Member
| Deliverable | Level 1 | Level 2 | Level 3 |
|---|---|---|---|
| Onboarding SOP | The member's documented process | Same (this IS the AI's reference) | Same |
| Roadmap template | Populated, member fills per-client fields | AI produces per-client version | AI produces per-client version |
| Action plan template | Populated, member fills per-client fields | AI produces per-client version | AI produces per-client version |
| Mutual responsibility template | Populated, member fills per-client fields | AI produces per-client version | AI produces per-client version |
| AI production prompts | — | Step-by-step prompts for Claude | Embedded in kit instruction file |
| Practice context file | — | Document Claude reads | Embedded in kit instruction file |
| Kit instruction file (CLAUDE.md-style) | — | — | Full instruction set Claude loads |
Level 1 is always produced. It's the foundation. Levels 2 and 3 build on it.
Production Status
| Member | Extraction | Level 1 | Level 2 | Level 3 | Notes |
|---|---|---|---|---|---|
| Rob Foncannon | Complete | Golden: 03a-tpc-golden-rob.html | Not built | Not built | Needs L1 SOP + L2/L3 kit files |
| Bev Stitely | Not started | — | — | — | Run extraction first |
| Tracy Beveridge | Not started | — | — | — | Run extraction first |
| Diane Gardner | Not started | — | — | — | Run extraction first |
| Linda Stapf | Not started | — | — | — | Run extraction first |
| Shelley Johnson | Not started | — | — | — | Run extraction first |
| Meryl Greenwald | Not started | — | — | — | Run extraction first |
Production order: Extraction → Level 1 (SOP + templates) → Level 2 (prompts + context file) → Level 3 (kit instruction file). Sequential — each level requires the previous one.
TPC-Specific Member Data
Member tool stacks
| Member | Practice management | Other tooling |
|---|---|---|
| Rob Foncannon | ATOM (portal-only enforcement) | QuickBooks Online, RC Reports, Claude, tax planning software |
| Bev Stitely | ATOM (strong workflow) + Drake (tax prep) | Additional stack TBD via extraction |
| Tracy Beveridge | Microsoft Access, migrating to Canopy or Financial Cents (ATOM ruled out 5/18) | Pro Series, Calendly |
| Diane Gardner | Portal-only (PMS TBD) | Mike Milan's Clear Path to Cash app, Keap (CRM) |
| Linda Stapf | Karbon | Claude ("Ted"), SmartVault, eMoney, Ping, ChatGPT |
| Shelley Johnson | TaxDome (with sealed-binder system) | Stack TBD via extraction |
| Meryl Greenwald | Hybrid paper-digital; exploring Financial Cents | Stack TBD via extraction |
Package-design notes
- Bev's institutional model — Excel table of services attached to engagement letter. Marked = included; unmarked = separately priced. October cutoff for tax-return inclusion. Gym-membership framing with 5% annual increase. Cleanup as separate project.
- Rob's lean model — One tier with named carve-outs. Investment management separate (series license pending). Up to three tax-planning strategies per year included.
- Diane's coaching model — Group + 1:1 + activation cycle defined; project work scoped separately.
Practice type classification
| Member | Practice type | Variant |
|---|---|---|
| Rob Foncannon | Lean tax practice | Solo practitioner with small team, one package tier |
| Bev Stitely | Institutional tax practice | Firm with productized packages, table-of-services model |
| Tracy Beveridge | Lean tax practice | Solo practitioner, no current packages (building) |
| Diane Gardner | Coaching practice | Group + 1:1, activation cycles |
| Linda Stapf | Lean tax/advisory practice | Solo with subcontractors |
| Shelley Johnson | Lean tax practice | Solo, carrying every role |
| Meryl Greenwald | Project-based practice | Hybrid — project-based with advisory elements |
Cohort-level reference data
All TPC member-specific proper nouns live in business-aos/cyp/tpc/tpc-reference-data.md and per-member files in business-aos/cyp/tpc/members/[name]/.
Relationship to Other Kits
kit-advisory-onboarding-v2/(universal) — this kit extends that one. Methodology, extraction, output intelligence, quality framework live there.kit-advisory-onboarding/+kit-tpc-advisory-onboarding/(v1) — predecessors.
Scoped kit v2 created 2026-05-20. Three-mode design added 2026-05-20. Methodology lives in kit-advisory-onboarding-v2/; this kit holds TPC-specific overlay and defines member deliverables.