Asset Zero · the plan

What ships, how we get there, and your part in it

The Human–AI Boundary Framework v1.0 freezes on 31 August 2026 — built from 2,796 recorded responses and validated by volunteers across three review cycles.

AI drafts, you judge · Every input gets a written answer · Rough consensus, named chair · Dissent is recorded, not erased · Named human sign-off.

What ships on 31 August

The six outputs promised to members, under their promised names — all living inside deliverable zero. Chips update as sections close.

The Framework v1.0

Living → v1.0 · 31 Aug

18 sections — a living site plus a citable snapshot PDF, frozen 31 August. Every other deliverable lives inside it.

Lifecycle AI Map

Draft (AI)

What the community said AI should do at each lifecycle stage, from the four webinars — the figure in §A3.

Member Case Library

Draft (AI)

The §C2 failure modes & lessons library: 13 consolidated modes with field prevalence, grown through August by volunteers' break-a-principle answers, reviewer cases and submitted case studies.

Decision Rights Matrix

In review (cycle 2)

§B3: 12 asset-management decision rows × AI level × what humans retain × named accountability, populated from the community's boundary votes. Closes cycle 2.

Boundary Questions Toolkit

In review (cycle 2)

§B4: the nine questions plus screening guidance keyed to asset criticality. Closes cycle 2.

Open Questions Register

Draft (AI)

§D1 — already live: OQ-01…19 in IETF-style states, each with what would settle it.

Recommendations for Members

Draft (AI)

§D2: the nine recommendations, prioritised by the community's final-pass vote.

Closed sections with published disposition logs

First close Mon 17 Aug

Every input answered in writing — your comment, what we did with it, and why. Links land here per cycle as sections close.

The provenance appendix (D4)

Draft (AI)

Participation stats, the disposition archive, and the evidence trail back to the recorded responses.

Named credits

Live — freezes 31 Aug

Working Group co-developers on the cover; reviewers on the contributors page; every participant acknowledged. Live on the snapshot now.

The boundary ledger

Live

The project's own AI governed under pre-registered levels — every action and denial public.

How we get there

AI drafts from the community's evidence; volunteers judge; the chair closes on rough consensus; dissent is recorded; named humans sign.

  1. Build

    Mon 3 – Fri 7 Aug

    **Stream 1 drafts the complete framework v0.5** (all 18 sections, provenance-tagged, from the evidence spine + W5B) · fact-check · Kai signs · **kickoff email goes Mon 3 or Tue 4 Aug** · volunteers confirm role + the August-hours consent (§4)

    Closes: v0.5 full draft

  2. Kickoff + Cycle 1

    Mon 10 – Fri 14 Aug

    **Kickoff call Mon 10, 12:00–12:30** [TBC] · cycle-1 packet lands same day · micro-asks close Wed 12 noon · **call Thu 13, 12:00–13:00** · dispositions + close Fri 14 · Reviewer packet 1 (Mon 10 → Fri 14)

    Closes: **A2 + B2 + B1** — one finding, seven principles + structure question, ladder & human roles → v0.6

  3. Cycle 2

    Mon 17 – Fri 21 Aug

    Packet Mon 17 · asks close Wed 19 noon · **call Thu 20, 12:00–13:00** · dispositions + close Fri 21 · Reviewer packet 2 (Mon 17 → Fri 21)

    Closes: **B3 + B4 + B5** — Decision Rights Matrix, boundary test, envelopes/trigger/stop → v0.7

  4. Cycle 3 — final pass

    Mon 24 – Fri 28 Aug

    **Whole-document last call** (async, IETF-style): 72-h objection window Mon 24 – Wed 26 on the full v0.7 + first-look C/D/A sections · **recommendations priority vote** (form, 15 min) · optional drop-in clinic Thu 27, 12:00 (attendance not required, not counted in anyone's budget) · dissent recorded · chair closes all sections Fri 28

    Closes: **C1–C5 + D1–D4 + A1/A3/A4** + whole-doc → v1.0-rc

  5. Freeze

    Mon 31 Aug

    **v1.0 tagged** · snapshot PDF assembled and published · participation stats + disposition archive into D4 · thank-you + credit confirmation to every volunteer

    Closes: **v1.0**

v0.5 → v0.6 (cycle 1) → v0.7 (cycle 2) → v1.0-rc (final pass) → v1.0 (31 Aug)

The one-week cycle rhythm, in full

| Day | What | Who | Effort | |---|---|---|---| | Mon | Packet out (brief + drafts + asks + reviewer case) | AI drafts · Kai approves & sends | — | | Mon–Wed noon | Micro-asks (2–3 structured, ≤40 min total) | Working Group | ≤40 min | | Wed noon | Response check: <60% → Kai's personal two-line nudge same afternoon | Kai | — | | Thu 12:00 | The call — contested points only, 60 min, rough-consensus test | WG + chair | 1 h | | Mon–Fri | Reviewer packet (one case + 3 questions) | Reviewers | ≤1 h | | Fri | Disposition log closed (AI drafts, chair adjudicates every row) · section status set · revised draft + log published | Chair + Kai | — | Missed a call? Recorded + 15-min catch-up pack; async input carries identical weight — **decisions close on the log, not in the room.** Rough consensus needs objections *addressed*, not attendance; participation stats publish transparently in D4 either way.

Your part

Working Group: ~4½ hours total, all in August (budget 5) — 30-min kickoff (Tue 11) + two cycles of short asks (≤40 min) and one 60-min call each (Fri 14, Thu 20) + a ~45-min final read and vote (wk of 24 Aug). Reviewer: ~2 hours — two 1-hour case packets. This concentrates the advertised ≤4 h/month into one bigger month and ends two months early — the kickoff email asks you to confirm you can wear it; async-only participation on whatever you can give still counts and is still credited. One browser link; no new tools — everything happens in Fulcrum at /members/tasks. The Volunteer Code of Conduct you agreed to at signup applies throughout.

Confirm your role in Fulcrum — the kickoff email's link lands on the RSVP: Working Group · Reviewer · both · async-only.

Sign in to RSVP — under three minutes, and your to-do list appears right here. Open to AMCouncil members; how credit works →

TierWhoWhere your name appears

Co-developer

Working Group membersNamed on the framework's cover page and in the snapshot's front matter

Reviewer

Volunteer reviewersNamed on the contributors page

Community contributor

Every webinar participantCollective acknowledgment (responses were anonymous and stay that way)

Step back at any time — the credit for cycles you contributed to stays. Credit follows participation, not attendance: answering the structured asks counts, and decisions close on the written log, not in the room.

The AI on this project

Your AI guide

In build + red-team — from 17 Aug

What it does

  • Walks you through your to-dos and deadlines.
  • Saves your answers only when you confirm them — through the same member door as the forms, under your identity.
  • Answers any project question, citing its sources.

What it can't do

  • Change framework text — drafting stops at L3, a hard cap.
  • Touch member data — it holds no credentials to any PII store.
  • Act unlogged — every action and refusal is a public ledger row.

AI synthesises the community's 2,796 responses, drafts sections with line-by-line provenance, and compiles the disposition logs. Humans set direction, judge content, run the calls, sign the sections and own the outcome. The project itself runs under the framework it is producing: the AI's permissions are declared in advance, every action and refusal is logged, and the log is published. A full AI-use statement appears in the framework's front matter.

The Human–AI Boundary Framework for Asset Management — to our knowledge the first of its kind for the sector — built from 2,796 community responses across four webinars (June–July 2026) and validated by you across three review cycles in August 2026 (11–14 · 17–21 · 24–28 Aug, whole-document final pass). Published as a living framework online plus a citable Version 1.0 snapshot, frozen 31 August 2026, with a Webinar 5 launch event to follow. It is community-developed guidance published by the Asset Management Council — practical guidance carrying the community's authority, not a formal standard.