Version. 0.1
Date. 2026-07-28
Status. Investigation. Develops the Operator's proposition recorded as Appendix A of the venture-diligence inspection report. No change request is implied; one adapter arc and one pilot are proposed for Operator decision.
Author. Claude.ai. Operator: Marvin Percival.
Origin. The Operator, mid-session 2026-07-28: rather than a fund using Loomworks to conduct diligence on a target, the fund makes FORAY adoption a condition of the term sheet — the target instruments its own operations, and the investor's view of the business is thereafter derived from actual business events rather than reconstructed from reporting. Captured by Claude Code as marked commentary (inspection report Appendix A); developed here now that its named prerequisite — the OVA reach-enforcement resolution — exists.
Grounding. Inspection report Appendix A; scoping-notes/loomworks-ova-reach-enforcement-scoping-note-v0_1 §§1–4; change-requests/cr-2026-157-ova-slice-1-non-member-contribution-pathway-v0_1; loomworks-vertical-venture-capital-structure-note-v0_1; loomworks-foray-ova-integration-strategy-investigation-v0_1; the FORAY Protocol (DUNIN7; universal transaction grammar; attestation at the engine substrate chokepoint at the moment of the act).
The proposition. Diligence looks at a company for weeks; the fund then holds it for years, seeing it through board decks — periodic, curated, self-reported, with a structural incentive to smooth. If FORAY adoption is a term-sheet condition, the target's operational systems emit attested events from the moment of close, and the investor's picture of the business is derived from what happened rather than assembled from what was reported. The reporting relationship changes from assertion to observation.
The strongest claim in it. FORAY is a universal transaction grammar. If every portfolio company's billing, payroll, and procurement events speak the same grammar, the metric-standardisation problem — every company reporting different numbers, the pain-point list's number five — is solved at the source instead of patched by mapping. Cross-portfolio questions become expressible against events, not against reconciled spreadsheets.
The honest counterweight. Adoption can be mandated; candour cannot. Section 6 takes this seriously and finds the answer is in what gets instrumented: systems, not people. A billing event does not lose candour under observation; an authored narrative does. The design keeps the two apart.
What is asked of the Operator. Three decisions at the end: confirm the hosting model recommendation (investor-hosted first), commission the FORAY-stream adapter as a named arc in the ingestion family, and decide whether the pilot design goes in front of a friendly fund. One counsel flag — material non-public information handling for funds with public-market exposure — is raised, not answered.
Conventional diligence and conventional portfolio reporting share one evidentiary basis: retrospective reconstruction from artifacts the company chose to assemble. The data room is a curated pile; the board deck is a periodic abstraction; both are authored about the business after the fact, by the party with the strongest interest in how it reads.
An event substrate inverts the basis. FORAY attestation fires at the engine chokepoint at the moment of the act — attestation is part of the transaction, not evidence assembled about it afterward. The record is contemporaneous, append-only, and generated by operations rather than authored by people. The investor reading it is reading what happened. The interesting claim is not better analysis of the same material; it is a different class of material.
This is the same principle the diligence vertical already carries — provenance on the claim, corrections preserved, machine origin marked — extended from the diligence file to the business itself.
The venture-diligence documents fixed a truth worth repeating: diligence is weeks; the hold is years. Everything the fund knows during the hold arrives through self-reporting on a quarterly cadence, and the smoothing incentive is structural — the people assembling the deck are the people the deck evaluates. When something goes wrong, the fund typically learns after the narrative has been managed.
Derivation from committed events changes three things. Latency: the fund's picture is current within the event stream's lag, not the board cadence. Selection: the fund's questions select from the record; the target's authors do not select for the fund. Trajectory: because the stream is append-only, "what did we know and when" has an answer during the hold, exactly as the diligence record gives it an answer during the deal — and at exit, the buyer's diligence inherits a company whose record is already evidence-grade, which is worth real money in the sale process.
The fund record vertical is where this lands: the fund's decade-long engagement gains, for instrumented portfolio companies, a source that is neither the deal record (frozen beliefs at entry) nor the board deck (managed narrative), but the operations themselves.
Pain point five in the source material: no universal reporting standard — one company reports annual recurring revenue and net revenue retention, another bookings and active seats, a marketplace reports gross merchandise volume and liquidity — and investors burn time normalising before any comparison means anything. The pain-points mapping honestly bucketed this as method work: recorded normalisation mappings, provenance-preserved, but mappings nonetheless.
FORAY dissolves the problem one level down. A universal transaction grammar means the revenue event, the payroll event, and the procurement event are the same shape at every instrumented portfolio company. Metrics become derivations over uniformly-shaped events rather than reconciliations of differently-shaped reports. The fund's cross-portfolio question — concentration, burn quality, revenue durability — runs against the grammar, and a new portfolio company joins the comparison set by adopting the grammar, not by mapping its spreadsheet.
No other tool in the diligence-tooling landscape can make this claim, because it requires the grammar to exist and the attestation to fire at the substrate — both of which are DUNIN7 protocol assets, not features to be copied.
Model 1 — investor-hosted (first, substrate-ready). The monitoring surface is the fund's: a per-portfolio-company operational engagement under the fund's organization, sitting beside the deal engagement it grew out of. The target's instrumented systems feed it through a FORAY-stream adapter (the engagement-as-universal-adapter pattern; Section 7 names the arc). The target's people, where human contribution is needed — context on an anomaly, a management explanation — contribute as credentialed non-members per CR-2026-157: attributed, held-until-admitted, reading nothing. The fund's members read; the fence does the rest. This is the Appendix A shape ("target contributes, investor reads") and it runs on what exists plus one adapter arc.
Model 2 — target-hosted (the maturity state). The target runs its own Loomworks engagements over its own operations — its record, its sovereignty — and the fund is granted governed reach into a restricted scope: an access-control list naming the fund, shaped by an external-account discipline on the target's side. This is the better long-term story (the target owns its record; multiple investors read the same source; the record survives the fund relationship) and it is honestly further out: it needs Slice 3's reach enforcement, OVA-proper credentials, and cross-organization reference — the federation direction. The investigation names it so Model 1 is built as a step toward it, not a dead end: the adapter, the grammar, and the shaping discipline all carry over; only the hosting flips.
Recommendation: Model 1 first. It ships on the current arc's outputs, it is the pilot shape, and nothing in it is thrown away at Model 2.
The proposition fails commercially if it is pitched as total visibility, and it should not be. What the fund sees is a shaped view: the term sheet negotiates a visibility schedule — which event families, at what granularity, with what lag — and the schedule is enforced where everything in this stack is enforced, at the substrate, not by the target's goodwill in assembling extracts. The target's commercially sensitive detail below the schedule never enters the fund's engagement; what enters is attested and complete within the negotiated scope.
This matters for the sale, in both directions. To the fund: complete within scope beats complete-in-theory, because it is verifiable. To the target: the schedule is a ceiling as well as a floor — the instrumented record proves the fund saw what was agreed and nothing more, which no email-me-a-spreadsheet regime can prove.
Appendix A's second caution, and the one that kills naive versions of this idea: an instrumented record that the investor reads changes what the target's people are willing to commit to the record. Mandating adoption does not mandate candour.
The design answer is a boundary, not a hope: instrument systems, keep authored content out of the instrumented claim. A billing event, a payroll run, a procurement transaction — these are generated by operations, and observation does not change what a billing system emits. The chilling effect is real but it applies to authored contributions — narratives, explanations, forward-looking commentary — and those stay where the diligence vertical already puts them: human contributions into the engagement, under the confidentiality constraints, shaped per reader, with the contributor told at the point of contribution exactly where their words can travel. The pitch never claims the instrumented stream captures sentiment, intent, or the conversation in the hallway. It claims the transactions are true, and that is the claim events can carry.
Two further realism points. The target gets standing value independent of the investor relationship — its own attested operational record, exit-ready, audit-ready — which converts adoption from pure imposition to partial self-interest; the term sheet supplies the activation energy, not the entire motive. And the fund's conduct is itself on the record: reads of the target's stream are attested too, which is a governance answer to the target's board when the question "who is watching what" gets asked.
One boundary stated plainly: this is not employee surveillance. The grammar describes transactions, not people's behavior; nothing in this proposition instruments communications, activity monitoring, or anything aimed at individuals. A fund that wants that is asking for a different product, and the answer is no.
One friendly fund, one portfolio company, one instrumented system, one quarter.
Instrument the revenue/billing event family only — the highest-value, lowest-sensitivity family, and the one every board deck claims to summarize. Run the stream into a Model 1 monitoring engagement under the negotiated schedule. At quarter end, produce the derived revenue picture from the engagement and set it beside the board deck's revenue section for the same period. Measure three things: divergence (where the deck and the derivation disagree, and why — smoothing, timing, error); latency (when the engagement knew what the deck eventually said); and the candour question observed rather than assumed (did anything about the target's behavior around the instrumented system change — the Section 6 boundary tested against reality).
The pilot's success condition is not "the deck was wrong." It is that the fund's partners, shown both artifacts, can articulate what the derived view is worth to them per year — which prices the offering — and that the target's chief financial officer can say what it cost them — which prices the adoption ask. Both numbers are things no document can produce; only the pilot can.
Not surveillance (Section 6's stated boundary). Not a claim that events capture judgment — evaluating people, strategy, and market timing stays human, exactly as the source material's own conclusion held. Not a replacement for the diligence vertical or the board relationship; it is the third element of one package: the deal engagement holds the diligence weeks, the fund record holds the fund's decade, and the instrumented stream holds the business's own account of itself between those two. And not a Loomworks feature: FORAY adoption by the target stands on the protocol, with Loomworks as the environment in which the fund reads it — the protocol triangle doing what it was drawn to do.
D-1 — Hosting model. Confirm Model 1 (investor-hosted) as the build-and-pilot shape, Model 2 preserved as the maturity direction. Recommendation: confirm.
D-2 — The adapter arc. Commission FORAY-stream adapter scoping as a named arc in the ingestion-interfaces family — the proposition's critical path. Recommendation: commission, sequenced after CR-2026-157 execution so the scoping session inherits a settled authorizer and credential substrate.
D-3 — The pilot. Decide whether Section 8's design goes in front of a friendly fund now (as a conversation, not a demo — the no-recall-demo rule and the adapter's non-existence both bar a demo) or waits for the adapter arc. Recommendation: the conversation can lead the build — the pilot design is itself a diligence instrument on the market's appetite, and a fund's reaction to Section 5's visibility-schedule framing is information worth having before the adapter is sized.
DUNIN7 — Done In Seven LLC — Miami, Florida Loomworks — FORAY operational visibility for investors — investigation — v0.1 — 2026-07-28