OptiArms

Carrier identity

The details a carrier reuses are what give it away.

Carrier identity theft is a paperwork crime with a data trail. Every phone, email, domain and address that arrives with a carrier is written down, and the ones that have been somewhere else before are what open a case.

The problem

What this is for

A hijacked MC number arrives with correct paperwork, because the paperwork is somebody else's. What it cannot bring is a history: the phone is new to that authority, the domain is weeks old, and the same details are sitting under two other carriers you booked last quarter. Nobody checks that by hand across a book of thousands.

The workflow

8 stages, and what runs where

Each stage names the module behind it. The chip says whether it runs on your network or ours.

  1. In your systems

    A carrier is added or changed

    Somebody sources a carrier, imports a sheet, or edits a contact in your TMS. That is the moment the details exist, and the moment they are worth reading.

  2. In your systems

    Your system posts the event

    One call, with the carrier, the contact details, and your own identifier for the record. Idempotent on that identifier, so a retry is not a second carrier.

  3. At intake

    The event is accepted or refused, in writing

    An unknown kind, a duplicate, an entity nobody is monitoring: each is recorded with its reason rather than dropped. The refusals are what the measured rate is later counted against.

  4. At intake

    The details join the ledger

    Every phone, email, domain, address and name is appended against the carrier, with where it came from and when. Nothing is overwritten, so the history is readable in both directions.

  5. On the platform

    The rules read the ledger, not just the event

    A value shared with another carrier, an authority younger than six months, contact details that changed in the last three weeks, a number that carries no line at all. Each rule that fires writes one sentence and attaches what it saw.

  6. On the platform

    Evidence accumulates on the carrier

    Findings from different rules pool as different witnesses; the same finding repeating is discounted rather than counted again. The result is a band, recomputed from the rows on the record.

  7. On the platform

    A case opens, or nothing does

    Above the floor your team set, a case opens and your roster is paged on the ladder you configured. Below it, the findings stay on the carrier and reach nobody.

  8. With a person

    Somebody decides, and it is recorded

    Confirmed, not fraud, or inconclusive, by a named person against the evidence in front of them. Nothing here decides, and nothing here tells the carrier anything.

What it runs on

Brokers, carriers, 3PLs, shippers, and factors, on the freight and logistics configuration. Verticals are settings over one platform, never separate products.

You supply the events

Carrier records, load tenders, correspondence, portal submissions. Posted from the systems that already hold them, over an API key you can revoke.

We supply the ledger and the rules

The identity history, the cross-reference across your own book, the published rules, and the case queue your team works.

Your system keeps the enforcement

A hold is a webhook asking your TMS to pause a load. What a hold means, and what releases it, stays a decision your own system makes.

Why this shape

What the design buys

The check runs on every carrier

Not the ones somebody had time to look at. A book of ten thousand is read the same way as a book of ten.

The reasons travel with the case

One sentence per finding, with the evidence under it. An analyst opens a case already knowing why it is there.

Your outcomes are the measure

The false-positive rate on your console is counted from verdicts your own people recorded, not from a number we published.

Limits

What it refuses

Stated here rather than found later. Each of these is enforced somewhere in the code, not only written on a page.

A shared value is not a finding

Dispatchers, factoring desks and answering services legitimately sit under many carriers. A value shared by too many stops linking anything and is offered to your allowlist instead.

A name is never evidence of a link

Two carriers with the same contact name are two carriers. Names are recorded for the timeline and never matched on.

It reads what you send it

A carrier sourced in a channel that never reaches the platform is a carrier it has no history for. Coverage of your own systems is the ceiling on what this can find.