# 16 · Make them insurable

*Part five — The proof and its absence*

---

Memo 9 hands the work its name:

> almost like the tagline is "make your agents insurable, which is quite an interesting angle, because it also allows us to do a nice market research

*Stated* — memo 9, verbatim. As positioning it is honest twice over: it places the work as *preparation for a market* rather than a product in one, which is true, since the market does not exist; and it does not require the market to exist before the work has value, which is the not-in-line position's commercial survival condition. But the memo's real cargo is the research programme underneath the tagline — because to say what makes an agent insurable you first have to establish what agent insurance *exists*:

> what insurance exists for agents? Like, can you insure Claude? Can you insure you know my Claude code session? Can you insure this session? And how much does it cost? And what does it cover?

*Stated.* *Can you insure this session* — asked, in a voice memo, about the class of session that transcribed the memo, processed it into doctrine, and wrote this book. The doctrine names what that question is:

> The survey is the first thing in this series that could be wrong in a way the world would correct.

*Stated* — doctrine 09. As of memo 9 the pivot had produced seventy-one decisions and no external evidence, and the count has only grown; every document in the folder is reasoning, checked against the corpus, entirely internal. The survey's answer exists outside the estate and nobody here has looked. For a body of work whose own rule is that a claim must be checkable, that matters more than any position in the preceding chapters — and this book's front matter carries it as position three because a reader who missed it has missed the epistemic status of everything else.

## The guard-rail, and the sentence nobody wants on a landing page

The positioning's hazard mirrors the folder's rule about levels:

> "Make your agents insurable" must not become "make your agents look insurable."

*Stated* — doctrine 09. Insurability is achieved by narrowing the delta or covering it — never by documenting it. And the honest counterpart, said out loud where a marketing document would bury it: **some placements should not be insurable.** A desktop agent under a user identity with network egress and no containment may be uninsurable at any level, and the useful output is that sentence. Finding one is a success of the method, not a failure of the customer. *Drawn.* This is chapter 8's badge hazard given its operational form, and the only structural defence the corpus has is the one it keeps reaching for: a programme that can output *no* is the only kind whose *yes* means anything.

## The mapping that upgrades a rule

The memo's inverse question — *what can you do that breaks the insurance?* — defines cover more sharply than cover does, and insurance has three classical answers. Two the estate already carries: **exclusions** are the mandate's prohibitions; **warranties** are the facts, failing as chapter 13 described. The third is the finding of part five:

> **A control declared on the card that the twin cannot find is not merely a weak input. It is the shape of material non-disclosure** — the thing that voids a policy rather than merely worsening a level

*Stated* — doctrine 09, GM-D73. Chapter 6 planted this quantity as a rating input; here it is promoted. Material non-disclosure — you did not disclose something that would have changed the terms — is the classical ground on which claims are refused, and the card-versus-twin gap has exactly its shape. Which changes what the gap is worth computing for: it has been an accuracy device; it is also **the single most consequential quantity an underwriter would want**, because it decides whether a claim gets paid. And the estate can compute it today, from two documents it already holds. *Drawn.* Whether any carrier or court would treat a computed card-versus-twin diff as legal non-disclosure is unknown and outside this estate's competence — the doctrine says so — but the direction of the finding survives the caveat: the pivot's most litigable quantity is one the measurement infrastructure already produces as a by-product.

## The survey is measure.py pointed at a market

The memo wants the market's pulse taken repeatedly — *as the market evolves, you can keep track of it* — and the doctrine's design move is to refuse to invent a methodology when the estate already has one:

> The survey is measure.py pointed at a market instead of a container

*Stated* — doctrine 09, GM-D74 — inheriting the tool's discipline unchanged: record what a product states publicly, never characterise what it privately offers, the same self-restraint that has the tool report presence and never contents; every finding carries its evidence class — a published price is observed, a policy wording is read, a vendor claim is documented; **unknown is never absent** — *we could not establish whether X covers agent placements* is the finding, not *X does not*; the survey is a floor, not a census; and it is dated, because an insurance observation goes stale fast, and a survey with no date is a claim about now that was true once.

And the one prediction is published in the only honest tense available:

> the survey will find no product insuring an agent placement as a unit

*Stated* — doctrine 09 §5, stated once so it can be falsified. If it holds, it is the strongest argument for the positioning; if it does not, the positioning changes — which is what evidence is for. The doctrine is explicit that predicting the survey's answers and publishing them as research would be the estate's cardinal sin, committed in the one document whose entire value is that somebody went and looked.

## Where the survey pays

*Drawn.* The commercial punchline is quieter than the epistemics and worth the chapter's last word. Every process the memo lists — web form, spreadsheet, Word document, API — collects evidence by questionnaire: the declared channel, exclusively. This estate produces measured evidence with provenance, which none of them asks for. Half an advantage — because an underwriter does not want a JSON document they cannot interpret. The gap is not producing evidence; it is producing evidence in a form an existing process can consume. So the survey's most valuable output is not prices or terms but **the format the market accepts** — because that format is the connector specification, and chapter 6's rule already governs it: the rendering must carry the evidence classes across, or a measured fact and a declared one arrive at the underwriter looking identical, and our own tooling will have manufactured the exact non-disclosure this chapter exists to detect.