The Evidence Ladder: Narrate Below the Line of Proof

The studio runs a public claims discipline built on six evidence tiers, from Possibility to Enduring Venture. The rule is one sentence: never narrate above the evidence line.

The Six Tiers

Every public claim the studio makes sits on one of six rungs, and each rung caps what may be said. The ladder runs:

Possibility → Experiment → Operational System → Live Product → Commercial Validation → Enduring Venture

The tiers have literal meanings, and the permitted language descends with them:

• Possibility: a technological opening we can state precisely. Language: "we believe," "this may become." No artifact is required; a clear reading of the frontier is enough.
• Experiment: something exists and is being tested. Language: "we are testing," "early signal." The artifact can be rough; the claim cannot be.
• Operational System: it runs on real inputs, produces real outputs, and has real operators. Language: "it operates." This is the rung where public claims inflate most often, so the label does its hardest work here.
• Live Product: deployed and in production use, carrying the obligations of a product. Language: "it is live."
• Commercial Validation: someone pays, or a durable external commitment exists. Language: "it earns." Nothing below this rung may be described as commercially proven.
• Enduring Venture: the work survives across time, leadership, and market cycles. Language: "it endures." This tier is earned over years and can be claimed only in retrospect.

The ordering matters less than what it enforces: a claim is legitimate only when the named tier is observable. An experiment narrated as a live product fails the rule even when every adjective is true.

Why Language Must Track Evidence

A public claim is a debt instrument. It promises the reader that reality, if inspected, will match the words. When inspection finds less, the debt defaults — and the default is never confined to the single claim. It reprices everything else the speaker says.

Two properties of the current environment sharpen this. First, the audience now includes machines. Answer engines and agents cross-check an entity's claims against every other surface it publishes, so inconsistency between what a studio says and what its labels show is machine-detectable — and detection degrades citation trust at the source. Second, the AI frontier rewards exaggeration with unusual short-term generosity, which means the incentive gradient points directly at the cliff. A written ladder is a precommitment against a known temptation.

The deeper reason is internal. A studio that lets language drift ahead of evidence corrupts its own instruments. Teams start optimizing for the story instead of the rung; milestones get redefined to fit announcements; eventually the gap between claims and systems becomes invisible from inside. The ladder keeps the instruments honest because the requirements of the next rung stay concrete.

What the Line Costs

The rule has real costs, and pretending otherwise would violate it:

• Slower announcements. Work that is real but unproven is described as real but unproven. In a market that rewards momentum theatre, "Operational Experiment" reads smaller than the work is.
• Refused upgrades. A venture sometimes sits between rungs. The ladder forbids borrowing the higher label while evidence catches up; the claim waits for the tier.
• Published ceilings. Every label doubles as a confession. Declaring a system an experiment tells a skeptical reader exactly how far the evidence goes — and hands critics the same map.

These costs are paid on purpose. The alternative — narrating above the line — is cheaper per announcement and ruinous per decade.

What the Line Buys

The return compounds. Every claim a reader can verify makes the next claim more credible; every labeled limit gives the strong claims weight when they are finally earned. A studio whose language is provably bounded by evidence can say "this works" and have the sentence mean something.

The ladder also works as a planning instrument. Each rung has exit criteria, which turns "how is it going" into a question with an observable answer. And it keeps the status vocabulary closed: on this site "Live Product" means deployed and in use, "Operational Experiment" means exactly that — Geolify carries the first label, Outflow the second, and neither borrows the other's rung.

The Operating Rule

The rule is one sentence: never narrate above the evidence line. Below the line a writer may say anything true — describe mechanisms, argue positions, report what the work has shown. Above it, nothing. Missing evidence is a do-not-publish condition, not a gap to be covered with confident prose.

What this buys in the end is the only asset that matters to a studio working a frontier: a record that can be audited against reality and survive the audit. See the first light. Build what endures. Let the claims trail the proof by exactly one honest rung.