{"slug": "i-handed-my-business-to-four-different-ais-today-two-of-them-invented-facts-here", "title": "I Handed My Business to Four Different AIs Today. Two of Them Invented Facts. Here's the System That Caught Them.", "summary": "A developer running a small e-commerce operation built a \"continuity file\" handover system and a structured packet-relay protocol for coordinating multiple frontier AI models as advisory seats, then stress-tested it with four full polls in which two of the models invented facts. The system grades every AI claim as RECEIPTS, CORROBORATED, or SELF-REPORTED and requires a VERIFY LIVE section before any answer is used, with the operator required to state back figures and stop for confirmation before acting.", "body_md": "**Seller disclosure, up front:** I sell a template kit version of the system described below — [The Continuity Kit, $39 on Gumroad](https://jacksonsyn.gumroad.com/l/kvrrck) — and I give away a [free skeleton edition](https://jacksonsyn.gumroad.com/l/mmoqza). 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.\n\nI 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:\n\n**The business only exists inside sessions. When a session ends, the next brain boots from nothing.**\n\nRules 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.\n\nSo 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.\n\nOne document, six sections:\n\nOur 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.\n\nA 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:\n\n```\nYou are stepping in as the operator of my business. Attached is my\ncontinuity file — the master handover document. Read it completely\nbefore doing anything else.\n\nFirst, do this and nothing else:\n\n1. STATE BACK to me, in plain language: what the business is, the\n   last verified figures WITH the dates the document gives them,\n   the current spend envelope and next gate, the open fronts and\n   their next actions, and the single points of failure.\n2. LIST anything you cannot verify, anything stale or contradictory,\n   and any access you would need from me.\n3. Then STOP. No messages, no purchases, no changes — until I\n   confirm your summary is right.\n```\n\nThree design decisions in there do all the work:\n\nOne 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:\n\n```\nPACKET:  ROLE (seat + \"advisor only, you cannot act\") · CHARTER (one\n         line per non-negotiable) · STATE (only the relevant\n         continuity lines, dated) · QUESTION (one) · DELIVERABLE\n\nRETURN:  ANSWER (direct, first) · EVIDENCE (graded per claim:\n         RECEIPTS / CORROBORATED / SELF-REPORTED / NONE) ·\n         VERIFY LIVE (every number/date/rule to check before use) ·\n         CHARTER CHECK · UNKNOWN\n```\n\nAnd one grading law over everything, AI output included:\n\n**RECEIPTS > CORROBORATED > SELF-REPORTED.**\n\nAn 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.\n\nToday 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.\n\nThe 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.\n\nBut 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**:\n\nEvery 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.\n\nThat'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.\n\nYou don't need the kit to start:\n\nWhere 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.\n\nIf you'd rather start from a worked system than a blank page:\n\nSame 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.", "url": "https://wpnews.pro/news/i-handed-my-business-to-four-different-ais-today-two-of-them-invented-facts-here", "canonical_source": "https://dev.to/sunnyops/i-handed-my-business-to-four-different-ais-today-two-of-them-invented-facts-heres-the-system-1b30", "published_at": "2026-11-06 21:08:30+00:00", "updated_at": "2026-11-06 21:23:53.952879+00:00", "lang": "en", "topics": ["ai-agents", "ai-tools", "large-language-models", "ai-safety"], "entities": ["Gumroad", "The Continuity Kit"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/i-handed-my-business-to-four-different-ais-today-two-of-them-invented-facts-here", "markdown": "https://wpnews.pro/news/i-handed-my-business-to-four-different-ais-today-two-of-them-invented-facts-here.md", "text": "https://wpnews.pro/news/i-handed-my-business-to-four-different-ais-today-two-of-them-invented-facts-here.txt", "jsonld": "https://wpnews.pro/news/i-handed-my-business-to-four-different-ais-today-two-of-them-invented-facts-here.jsonld"}}