Source: frameworks/kit-advisory-onboarding-v2/00-advisory-onboarding-v2-start-here.md

00 — START HERE: Advisory Onboarding Kit v2

Level 3 kit design. This kit operates at Kathryn's maturity level — adaptive, self-correcting, generative. The AI makes decisions, resolves gaps, and only escalates true judgment calls. This is not a checklist that AI runs — it's a capability the AI owns.

What This Is

A kit that produces the client-facing onboarding artifacts for an advisory engagement — the roadmap, the action plan, and the mutual-responsibility document. Built from two inputs: a universal process (already extracted from cohort sessions) and per-member offer data (gathered by AI, not by Kathryn live).

Audience (kit user): The practice owner (advisor) — TPC cohort members initially; Practice Builders cohort and beyond.

Audience (output): The advisory client. The roadmap and supporting artifacts are client-facing.

Format: HTML for the roadmap (visual, client-facing, shared); Markdown for the supporting artifacts — populated with member-specific data, not bracket-filled templates.

Lifecycle: One kit run per practice. The output becomes the advisor's living document with their clients — it leaves the kit once delivered.

What Changed from v1

Two things changed. One is the input design. The other is the kit's own maturity level.

Input design fix

v1 required a 40-minute live interview with Kathryn for every member. Wrong — the universal process was already captured through 4 weeks of Monday sessions. Per-member offer data can be gathered by AI. v2 uses AI extraction (file 06) instead of Kathryn interview.

Kit maturity upgrade

v1 was a Level 1 kit: documented steps, checklists, templates. AI followed instructions. v2 is a Level 3 kit: the AI adapts to the practice, resolves gaps during production, draws on cross-member intelligence, and only escalates genuine judgment calls. The difference:

Level 1 Kit	Level 3 Kit
Extraction	Fixed questionnaire, 11 questions in order	Adaptive conversation, gets to the same data through natural dialogue
Output	Fill in a template with sections in fixed order	Build artifacts that reflect this specific practice — structure adapts to the practice model
Quality	Checklist of pass/fail boxes	Self-correcting framework — fix what you can, escalate what you can't
Cross-member	None — each run is independent	Uses what worked (and what didn't) from prior runs to improve the current one
Gaps	Flagged for Kathryn to resolve	AI proposes solutions based on what similar practices do — Kathryn only sees true judgment calls

This is the standard going forward. Every kit Kathryn builds operates at her maturity level, not the member's. Members receive Level 1/2/3 outputs. The kit itself operates at Level 3.

What This Produces

Deliverable	Format	Audience	Filename pattern
The roadmap	HTML	Client + advisor + team	`[practice]-advisory-roadmap.html`
Action plan	Markdown	Advisor populates per client	`[practice]-action-plan-template.md`
Mutual-responsibility doc	Markdown	Advisor populates per client	`[practice]-mutual-responsibility-template.md`

All three are produced as FINISHED artifacts with member-specific data filled in — not bracket-filled templates waiting for manual population.

Per-practice variants of all three live in the scoped cohort kit (e.g., kit-tpc-advisory-onboarding-v2/03a-tpc-golden-rob.html), not in this universal kit.

Operating Modes

Mode	Trigger	What It Produces
Mode 1: Create	A practice owner is ready to have their onboarding built. AI extraction completed.	The roadmap + action plan + mutual-responsibility doc, populated with their data.
Mode 2: Improve	After running the kit, output changed by hand, QC found a gap, or process needs adjustment.	Updated kit files (golden, output skill, quality checklist). See file 05 Mode 2.

No Update mode. Once a practice owner has their roadmap, it's their living document — the kit doesn't keep updating it.

The Two Inputs

Input A: Universal Process (ALREADY CAPTURED)

The Monday extraction sessions produced the methodology. This input is complete. The kit reads it; it does not re-extract it.

What it contains:

The five spine elements (pre-schedule the cadence, living document, mutual responsibility, first win, welcome experience)
The package-design layer (what's included vs. separate)
Design constraints (human-checkpoint requirement, personal-first deployment, stop-sending-forms, confidentiality posture, special-sauce protection)
Locked vocabulary (trigger, onboarding, pre-onboarding, roadmap, first win, cadence, completion gate, etc.)

Where it lives: file 01 of this kit (context), synthesized from the 4 weeks of extraction.

Input B: Per-Member Offer Data (AI-GATHERED)

A structured AI-guided extraction where the member answers questions about their specific offer. This replaces the 40-minute Kathryn interview. See file 06 for the full extraction protocol.

What it captures:

Package name and structure
What's included, what's separate
Pricing structure (boundaries only — not client-facing dollar amounts)
Team roles in delivery
Tools the client interacts with
Meeting cadence shape
First-win candidate
Welcome experience design
Completion gate criteria

Who runs it: Claude with the member (no Kathryn required). Can be done asynchronously, in a Thursday co-working session, or in a standalone Claude session.

Time: 10-15 minutes.

Maturity Model Integration

Every output artifact addresses three levels:

Level 1 (Structure): The documented process. The SOP version of the onboarding — what happens, when, who does what. Every member gets this regardless of AI interest.
Level 2 (Streamline): AI handles the repetitive parts. Draft communications, scheduling sequences, document collection follow-ups, progress tracking.
Level 3 (Delegate): AI does the heavy lifting end-to-end. The human reviews, approves, redirects. The onboarding runs itself; the advisor supervises.

Members self-select their depth. The kit produces all three layers in every run.

What This Does NOT Do

Does NOT require Kathryn to personally extract per-member data. The AI-guided extraction (file 06) handles this.
Does NOT re-extract the universal process. That work is done. The kit reads the synthesis; it does not re-run Monday sessions.
Does NOT bake one member's proprietary content into the kit's spine. The kit captures universal structure; tactical edge stays with the practice.
Does NOT produce pre-onboarding artifacts (discovery, presentation, signing process). Pre-onboarding is everything BEFORE the trigger; this kit produces what comes AFTER.
Does NOT produce bracket-filled templates. The output skill (file 05) produces FINISHED artifacts with member-specific data filled in.

Critical Rule — Client Perspective Test

Every section of every output produced by this kit must pass the client perspective test:

"If the client read this cold, with no advisor present, would it (a) make sense, (b) feel respectful of their time and intelligence, (c) tell them what to expect, what they're committing to, and what they'll get?"

If any section fails any of those three, fix it before delivering. This rule is non-negotiable.

File Inventory

#	File	What It Is
00	`00-advisory-onboarding-v2-start-here.md`	This file — orientation, what changed, modes, file inventory
01	`01-advisory-onboarding-v2-context.md`	Inputs (A + B), definitions, validation rules, patterns, special-sauce protection
02	`02-advisory-onboarding-v2-terminology.md`	Locked vocabulary — preserved from v1
03	`03-advisory-onboarding-v2-golden-example.html`	Rob's roadmap — complete, no gap flags, welcome experience and completion gate filled
04	`04-advisory-onboarding-v2-quality.md`	Quality FRAMEWORK — self-correcting, fix what you can, escalate what you can't
05	`05-advisory-onboarding-v2-output-skill.md`	Output INTELLIGENCE — generative production, adapts to practice model, self-QCs during build
06	`06-advisory-onboarding-v2-ai-extraction.md`	Extraction CONVERSATION — adaptive, cross-member intelligence, proposes solutions for gaps

Total: 7 files. Same count as v1 — file 06 changes from interview protocol to AI extraction protocol.

Relationship to Other Kits

kit-advisory-onboarding/ (v1) — predecessor. v2 replaces v1's input design. v1's methodology, patterns, and vocabulary are preserved. v1's interview protocol (file 06) is replaced by the AI extraction.
kit-tpc-advisory-onboarding-v2/ (scoped cohort overlay) — holds the actual TPC member goldens. References this kit for all methodology; does not duplicate.
kit-builder/ — produced v1 via Mode 1. v2 was built directly from the corrected design identified in the TPC evolution doc.

Self-Improvement Loop

After running this kit, ask:

Did I change anything in the output by hand? → Update file 03 (golden) + file 05 (output skill) via Mode 2.
Did QC miss something I caught? → Update file 04 (quality) — add the check + add to common failure modes.
Is there something the kit should do that it doesn't? → Update file 05 (output skill) or file 06 (AI extraction).
Did the AI extraction miss a data point the output skill needed? → Update file 06 (extraction) + file 01 (context, Input B definition).

Every manual fix becomes a kit fix. If you fixed it by hand, tell the kit so you never fix it by hand again.

File Location

advisory-os-vault/content/frameworks/kit-advisory-onboarding-v2/
    00-advisory-onboarding-v2-start-here.md          (this file)
    01-advisory-onboarding-v2-context.md
    02-advisory-onboarding-v2-terminology.md
    03-advisory-onboarding-v2-golden-example.html
    04-advisory-onboarding-v2-quality.md
    05-advisory-onboarding-v2-output-skill.md
    06-advisory-onboarding-v2-ai-extraction.md

Kit v2 created 2026-05-20. Level 3 upgrade 2026-05-20. Two changes from v1: (1) AI handles per-member extraction, not Kathryn. (2) Kit operates at Level 3 — adaptive, self-correcting, generative. This is the standard for all kits going forward.