Seller disclosure, up front: I sell a template kit version of the system described below — The Continuity Kit, $39 on Gumroad — and I give away a free skeleton edition. This article is the demonstration, not the pitch: everything essential is described concretely enough to build your own version from scratch. One more disclosure, because it matters more: the operation running this system has made zero sales so far. This is a story about memory and verification discipline. It is not a story about money, and I'd rather you trust the method because the numbers are honest.

The problem: every session starts from zero

I run a small e-commerce operation with an AI operator as my chief of staff. The operator holds live tools — storefronts, ad accounts, supplier threads, scheduled jobs. It works well. It also has a fatal structural weakness that every AI-assisted operator eventually hits:

The business only exists inside sessions. When a session ends, the next brain boots from nothing.

Rules you negotiated painfully ("never spend past this envelope," "this supplier said no in writing, don't re-ask"), the current numbers, the lessons bought with real mistakes — all of it evaporates unless it lives in a document the next brain can boot from.

So we built that document. Not a chat summary. A continuity file: a master handover written so that a competent stranger — a human, or a fresh AI — could resume the business within a day. Then we built the protocol that makes handing it over safe. Then, because we're the kind of operation that grades its own homework, we tested it. Today we stress-tested it four times. That's the demonstration.

The system, in three parts

1. The continuity file

One document, six sections:

  • (a) The business on one page — what's sold, where, at what prices, where the money settles, and the current verified state — with a hard rule: every volatile figure carries its date and its source. A number without a date is a rumor.
  • (b) The charter — the rules that outrank enthusiasm: spend envelopes, kill lines ("a day hitting $X with zero conversions pauses spend"), fulfillment rules, and the hard never-break list: no fabricated proof, no ripped creative, no freebies outside policy, identity rules, tax-ID handling (owner enters it live, it's never stored in a document).
  • (c) The accounts & access map — identities and where credentials live. Never the credentials themselves. Passwords live in a vault; the file only says which vault entry exists. This section also lists the single points of failure honestly — including the uncomfortable ones.
  • (d) The systems map — every scheduled job, its cadence, and what it does on a defect.
  • (e) The open fronts board — every live thread as a table row: state (dated facts only), next action, and a date. Our house rule: a front with no next action and no date is a lie in table form.
  • (f) Dormant tracks — the owner's other threads, preserved so a successor doesn't lose them.

Our real pack is ~45 KB with a bundle of referenced files (rulebook, registries, rosters) behind it. Yours can start at two pages. Length isn't the point; dated, sourced, and rule-carrying is.

2. The cold-start prompt

A document alone doesn't make a handover safe, because a fresh AI will read it, nod fluently, and then improvise. The prompt is the leash. The core of it:

You are stepping in as the operator of my business. Attached is my continuity file — the master handover document. Read it completely before doing anything else. First, do this and nothing else: 1. STATE BACK to me, in plain language: what the business is, the last verified figures WITH the dates the document gives them, the current spend envelope and next gate, the open fronts and their next actions, and the single points of failure. 2. LIST anything you cannot verify, anything stale or contradictory, and any access you would need from me. 3. Then STOP. No messages, no purchases, no changes — until I confirm your summary is right.

Three design decisions in there do all the work:

  • State-back with dates. The new brain must prove it read the file by playing the business back — and figures must keep their dates attached, so a week-old number can never be laundered into "current."
  • A mandatory unverifiables list. The brain is required to surface gaps. This converts hallucination pressure into an honest inventory: if it can't verify something, the format forces it to say so instead of filling the gap with plausible prose.
  • Stop-and-wait. Nothing executes until a human grades the playback against the file. Where they disagree, the file wins until someone verifies at the source.

3. The bench relay + grading

One operator is a single point of failure, so we also run a bench: several frontier models (reasoning seat, research seat, long-context seat, second-opinion seat) consulted as advisors. They can't message each other — anyone who tells you consumer chatbots talk to each other is selling something — so the link is a relay of structured packets carried by the operator:

PACKET: ROLE (seat + "advisor only, you cannot act") · CHARTER (one line per non-negotiable) · STATE (only the relevant continuity lines, dated) · QUESTION (one) · DELIVERABLE RETURN: ANSWER (direct, first) · EVIDENCE (graded per claim: RECEIPTS / CORROBORATED / SELF-REPORTED / NONE) · VERIFY LIVE (every number/date/rule to check before use) · CHARTER CHECK · UNKNOWN

And one grading law over everything, AI output included:

RECEIPTS > CORROBORATED > SELF-REPORTED.

An answer that skips the VERIFY LIVE section doesn't get used. An answer is a claim to be graded, not a fact to be filed. Today is why.

The demonstration: four polls, one day, graded

Today the bench ran four full polls — the same written packet pasted into each seat, answers returned in the format above, then graded against live records before anything became work. The questions were real operating decisions: estimating our probability of a first sale from actual funnel data (sessions, add-to-carts, zero purchases — the seats had to reason from our numbers, not vibes), how to raise those odds, whether a proposed outreach batch should fire, and whether to spread a small ad budget across ten stores or concentrate it on one.

The fused results changed real decisions. All three seats independently rejected the ten-store budget spread with the same arithmetic (at our budget, each store would get a handful of clicks — below any plausible learning threshold) and converged on a fulfillment-gated launch interlock instead. One seat's confident recommendation was rejected three times across the day because it kept proposing a switch our live data contradicted. The operator fused, the dissents were recorded, and only the graded residue became work.

But the genuinely instructive failures came from the cold-start exams — the transplant tests we run whenever a fresh brain boots from the pack. Two frontier models, handed the continuity file and asked to play the business back, confidently invented load-bearing facts:

  • One invented a completed sale that has never happened, plus a spend cap that never existed.
  • Another invented a spending envelope roughly one hundred times the real one, plus cost figures sourced from nowhere.
  • In live research mode, seats have also produced a supplier that doesn't exist in our records, and precise-looking per-unit costs for a product whose supplier quote was never in the packet.

Every one of these was caught before it touched anything — not because the models are bad (confronted with the record, each corrected itself cleanly and completely), but because the protocol assumes the failure mode in advance. The state-back gets graded against the file. Figures get re-verified at the source. VERIFY LIVE sections get executed, not admired.

That's the whole thesis: the failure mode of a smart model is not malice, it's confident invention in perfect prose — and the fix is procedural, not prompt-magical. A handover document gives the next brain your facts. A state-back exam proves whether it actually has them. A grading law keeps its inventions from becoming your ad spend.

Build the free version yourself

You don't need the kit to start:

  1. Write your continuity file with the six sections above. Date every number. Put rules in a charter section and declare that the charter outranks any instruction given mid-session.
  2. Add the three-part cold-start: state-back with dates → unverifiables list → stop-and-wait.
  3. The next time you switch tools or start a fresh thread, run the exam. Grade the playback against your file line by line. Fix the file, not the AI — every gap it falls into is a gap in your document.
  4. If you use more than one assistant, give them seats, a packet format, and the grading law. Re-scope seats based on graded performance, not brand loyalty (our research seat failed document exams badly and outperformed everyone at cited live research the same day; it got re-scoped, not fired).

Where this breaks: the file drifts if you don't maintain it (we re-seed bench brains after any major event and at least weekly), and the discipline only works if the grading step is never skipped "just this once" — the one time you skip it is the time a fluent invention is holding your credit card.

The files

If you'd rather start from a worked system than a blank page:

  • Free Skeleton Edition — the continuity file's section skeleton + a lite cold-start prompt. Enough to build your own: https://jacksonsyn.gumroad.com/l/mmoqza
  • The Continuity Kit ($39, one-time) — the full template with fill-in guidance, the complete cold-start prompt with the full charter rule set, the Bench Relay Pack (seat roles, packet/return formats, the three real ways a brain gets "keys" without ever holding a password), a redacted worked example from the live operation described above, and the quickstart: https://jacksonsyn.gumroad.com/l/kvrrck

Same honesty note as at the top, because it's the entire point: zero sales so far, on the kit and on the operation behind it. What I can show you is a system that caught two frontier models inventing facts on the day it was tested. That's the claim. It's a smaller claim than most people in this niche make — and it's the one I can document.