Automation stops where proof stops.
LENTRA is not bought by a developer looking for cleaner data. It is bought where a machine decision has consequences and a human name is on the outcome — and where automation was held back precisely because provenance could not be proven.
Where the cost of being wrong is already priced in.
Illustrative pressure index, not measured data: our read of how strongly each segment already needs provable inputs.
A decision is challenged months later and the input cannot be reproduced — only described.
The exact observation, its time, its vantage point and its permitted use are produced as a signed record.
Underwriting or claims rest on an external signal whose origin cannot survive a dispute.
Every signal used in a decision carries provenance strong enough to be filed alongside the policy.
A takedown or legal action depends on what was seen, when and from where — and the evidence is a screenshot.
Observations are signed at capture, with egress context, so counsel receives evidence rather than a claim.
Machine-set prices cannot be explained per market and per moment when a regulator or partner asks.
Each price move is anchored to the observations that caused it, with the market frame attached.
Enforcement against a seller or a listing is only as strong as the record behind it.
Actions cite verifiable observations, which shortens appeals and holds up under review.
An autonomous agent acts on the open web and every action becomes undocumented liability.
The agent's inputs are passported, so its actions can be audited afterwards rather than trusted in advance.
Buyers increasingly ask for provenance the provider cannot express in any standard way.
Emitting compatible passports turns the provenance question from a negotiation into a field.