01 — CONTEXT: Advisory Onboarding Kit
Input definitions, validation rules, the patterns the kit enforces, the special-sauce protection rule, and the audience boundary.
Mode 1 Inputs — Create
| Input | Required | Example / Source | Used For |
|---|---|---|---|
| Interview output (structural capture) | Yes | Produced by running file 06 protocol with the practice owner | Trigger, step sequence, first-win mechanic, cadence, delegation, real-vs-aspirational separation. The kit's primary substrate. |
| Practice reference data | Yes | Member-specific reference file (e.g., members/rob/rob-foncannon-reference-data.md) + cohort-level reference data (tpc-reference-data.md) | Proper nouns, team names, tool names, firm specifics. Overrides transcript renderings. |
| The five patterns + package-design layer | Yes | This file (below) — extracted from cyp/tpc/builds/advisory-onboarding-patterns-2026-05-04.md + 5/18 package-design addition | The universal spine the roadmap must address. |
| The cohort decisions table | Yes | cyp/tpc/drafts/kit-advisory-onboarding-build-bundle.md (Decisions made section) | Locked vocabulary, locked design constraints, locked patterns. |
| Audience boundary | Yes | This file (below) | What stays in (client-facing) vs. what stays out (advisor-internal). |
Validation Rules — Mode 1
- The interview output must come from a live interview. Solo recordings, written submissions, and paragraph-form inputs do NOT meet the bar. They produce patterns-input, not golden-input. If only solo material is available, route to the interview protocol (file 06) before running this kit.
- The practice reference data must be current. Tool stack, team names, renewal dates need to be accurate at the time of the build. Stale reference data leaks transcript-misrendered proper nouns (e.g., "Adam" not ATOM, "Carbon" not Karbon) into the client-facing output.
- The trigger must be locked before any post-trigger steps get built. The roadmap describes what happens AFTER the trigger. If the trigger isn't defined (signed agreement + payment received, by default), the rest of the roadmap floats. File 06 interview question #1 covers this.
- The package-design layer must be specified for institutional practices. Institutional practices (with productized packages, like Bev's Saunders Tax) need the package-design layer addressed upstream of the trigger — what's in the package vs. what's a separate project. Lean practices can use a simplified version. See "Package-design layer" below.
- Special-sauce protection is enforced at content-write time, not at output-review time. When building the roadmap, every element either traces to a universal pattern (file 01) or to the practice's own variant. No element of the roadmap requires another practice's proprietary content to function.
Input Priority Hierarchy
When inputs conflict or are ambiguous:
- Practice reference data wins for proper nouns (names, tools, firm details). Overrides transcript renderings, cohort data, and any other source.
- Interview output wins for process detail (trigger, sequence, cadence, first-win, completion gate).
- The five patterns + package-design layer win for universal spine elements (what every onboarding needs).
- Cohort decisions table wins for terminology (what "onboarding" means, what "trigger" means).
- Existing vault conventions win for file format, kit structure, and HTML styling.
The Five Patterns + Package-Design Layer
Extracted from 4 weeks of cohort extraction (4/27, 5/4, 5/11, 5/18) plus three practice-owner walkthroughs (one live interview + two solo submissions). These are the universal spine. Every roadmap the kit produces must address them.
1. Pre-schedule the cadence
Book the year's advisory meetings before the work starts — at onboarding, not as the year unfolds. First-in-line scheduling is the jankiness; pre-scheduling is the spine.
- Lean tax-practice variant: seasonal cadence anchored to tax/fiscal year — kickoff, mid-year check-in, fall working session, year-end review + next-year setup
- Institutional tax-practice variant: seasonal cadence + cleanup-as-separate-project carved out as a discrete pre-engagement
- Coaching-practice variant: activation-then-steady — weekly for 2-4 weeks, then monthly
- Flex point: the cadence shape (frequency, timing, anchor events) flexes by practice
- Spine: the year is on the calendar before work starts
2. Living document, not static PDF
The artifact format is a working document the firm owns and tracks, visible to client + firm + assistant. The firm drives it. The client contributes to it but doesn't own the tracking.
- Failure mode this prevents: PDFs filled out once, lost in browser sessions, never updated, never tracked. Linda's 25-year frustration with form-based onboarding.
- Spine: the roadmap is shared, updateable, and lives somewhere both sides can see.
3. Mutual responsibility, in writing, signed off
Every onboarding has a written, signed document specifying what the firm does, what the client does, by when. Format flexes; commitment doesn't.
- Lean variant: one-page action plan, "what I'm working on / what I need you to work on" with dates
- Coaching variant: formal signed mutual-expectations doc with named consequences for non-engagement
- Institutional variant: client expectations document covering firm's plate vs. client's plate, integrated with the engagement letter
- Spine: the document exists, both sides have seen it, both sides have signed/acknowledged it
4. First win, fast, measurable in dollars
A concrete, dollar-measurable result the client gets within 30 days of the trigger. Identified at kickoff; delivered within 30 days. Anything longer than 30 days loses urgency.
- Lean tax-practice variant: reduce the next quarterly estimated tax payment via a specific tax-planning strategy; client sees the dollar amount of the reduction at kickoff; execution within 30 days
- Coaching-practice variant: identify a cash leak via a diagnostic tool; client commits to closing one; deliver within 2 weeks
- Flex point: win type (tax savings, cash-leak recovery, fee-recovery, other)
- Spine: fast, measurable, the client can point to it. They feel the engagement justifies itself before steady-state kicks in.
5. Welcome experience > logistics
The first thing a new client receives is NOT a form to fill out. The welcome experience fills the space between "I'm a client" and "we're working together" with something that reinforces the decision.
- Variants: welcome video, welcome letter, gift, pre-call questionnaire (Diane), or some combination
- Spine: the first touchpoint signals what they just bought before the logistics begin
6. Package-design layer (added 5/18)
What's in the advisory package vs. what's a separate project. This is upstream of the trigger — without it, the rest of the onboarding builds on undefined scope.
- Institutional model: table of services attached to the engagement letter. Only marked services are included. Unmarked services are listed but priced separately. Calendar-cutoff rules for inclusion of seasonal deliverables (e.g., the first tax return). Cleanup-as-separate-project pattern.
- Lean model: simpler — one package tier with named carve-outs (e.g., investment management as a separate commission arrangement)
- Coaching model: group + 1:1 + activation cycle defined; project work scoped separately
- Spine: what's in the package is named. What's a separate project is named. The line is drawn before the engagement starts.
Special-Sauce Protection Rule
This kit serves multiple practitioners. The kit's spine is universal; member-specific competitive edge stays with the member. Without this rule, careful members hold back during extraction, and the kit ends up thin.
When building the roadmap:
- Capture universal patterns as the spine (e.g., "first win within 30 days, dollar-measurable")
- Capture member-specific tactics as variants noted with attribution (e.g., "the lean tax-practice variant: reduce quarterly estimated tax payment via a named tax-planning strategy") — in client-facing output, attribute to the practice using the kit; in cohort-internal documentation, attribute to the source practice
- Never make the kit's spine depend on another practice's proprietary content
Test: Does any element of the spine REQUIRE proprietary content from a specific practice to function? If yes, generalize the spine — the proprietary content moves to a variant.
See cyp/tpc/lessons-learned/lesson-special-sauce-protection-2026-05-19.md for full rationale.
Audience Boundary — In vs. Out
The interview output contains advisor-internal data. The roadmap is client-facing. Filter:
| Stays IN the roadmap (client-facing) | Stays OUT of the roadmap (advisor-internal) |
|---|---|
| The cadence — months, dates, meeting purposes | Pricing detail, package architecture decisions |
| The first-win identification + dollar amount | The advisor's 1-to-10 success-rate self-assessment |
| What the client will do, what the firm will do, by when | Failure modes the advisor named from prior clients |
| Named artifacts the client interacts with (roadmap itself, action plan, mutual responsibility doc) | The advisor's emotional friction about deploying to existing clients |
| Tools the client interacts with (QBO, client portal, scheduling link) | The advisor's internal PMS workflow (ATOM, Karbon back-end) |
| The completion gate language ("onboarding is complete when...") | Calibration notes for the next interview / kit run |
Default direction: when in doubt, exclude from the client-facing output. The advisor can always add it during their delivery; the kit can't unsay something it included by mistake.
Design Constraints
These are non-negotiable. Surfaced across 4 weeks of cohort extraction.
| Constraint | Source | What this means for the kit |
|---|---|---|
| Human-checkpoint requirement | Bev 5/4 — "I don't want it to update that document just from transcripts" | The kit cannot prescribe auto-updates to client records or reference data without a practitioner-review step. |
| Personal first, team later | Linda 5/4 deployment preference | The kit's rollout guidance supports solo-practitioner adoption before team adoption. The roadmap can be produced and used by an advisor alone, without team buy-in as a precondition. |
| Stop sending forms | Tracy 5/4 — clients don't know what they don't know | Easy data via form; real data via conversation. The kit's interview protocol IS this principle at the input layer. |
| Confidentiality posture | Bev 4/27 — "never give Claude a Social Security number" | The kit's roadmap and supporting artifacts must specify what NEVER touches AI tooling. SSNs, EINs, account numbers, and other identifying financial data stay in human-only channels. |
| Special-sauce protection | This session (2026-05-19) | See above. |
Gap Protocol
If any of the following are missing or insufficient, the kit cannot produce a golden-quality output. Stop and resolve before continuing.
| Missing | Impact | Resolution |
|---|---|---|
| Live interview (only solo recording available) | Output will be paragraph-form, not step-ordered. Patterns visible, structure absent. | Run file 06 interview protocol before running this kit. |
| Practice reference data | Proper nouns leak as transcript renderings into client-facing output. | Pause kit run. Populate or update the member's reference data file. |
| Trigger not locked | Post-trigger steps have no anchor. | Re-run interview question #1 (file 06). Lock the trigger before continuing. |
| Package-design layer not specified (institutional practice) | Scope creep risk; client confusion about what's included. | Add upstream package-design step OR document why this practice's package is bundled-everything (rare). |
| Completion gate not specified | The roadmap has no endpoint. The advisor and client can't agree when onboarding is done. | Define the completion gate criteria with the advisor. Default candidates: technical onboarding done + seasonal meetings booked + mutual responsibility signed. |
Pre-Build Validation Gate
Before producing any output, confirm:
- [ ] Interview output is structural capture from a live interview, not a solo recording
- [ ] Practice reference data is current (last updated within 30 days)
- [ ] Trigger is named and locked
- [ ] Package-design layer is specified (or explicitly waived for lean practices with bundled offerings)
- [ ] Completion gate is named
- [ ] Special-sauce protection has been applied to the captured content
- [ ] Audience boundary has been reviewed — advisor-internal items moved out