What Runink FACE Is Built For

The Problem Runink FACE Is Built Around

Every job above has the same shape. The facts you need are already recorded somewhere in your business. They sit in four systems, in four formats, and joining them up takes a morning that nobody has.

So the claim expires. The container is opened warm. The order goes out by air. Not because anyone made a bad call, but because nobody had the time to reach the point where a call could be made.

What Runink FACE Does About It

FACE is built to run its checks against your own records overnight, so that the morning starts with a short, ranked list of things that happened, with the records attached.

Each item is meant to be a proposed action, not an alert. The claim comes with the receipt, the reading, the rate and the deadline, and a drafted letter. The temperature drift comes with the container, the customer and a drafted rerouting.

A proposed action waits in that queue as a record of its own. Approving it is the step that carries it out, and the approval is written down with the name of the person who gave it and what they decided. An action nobody approves is an action that has not been sent.

Note what that is and is not. It is a queue whose items move because a named person moved them, and a record of who moved them — not a blanket gate somewhere in the system that inspects everything else your business does. Where part of a drafted action cannot be carried out, what comes back says so rather than reporting it done.

Approving is designed to end the work rather than start it. The reply names what went and what did not — a mail connector that is not configured comes back as a named skipped step, with its reason, on every reply. Later, when someone asks why a claim was filed or an entry was held, the answer comes from the record.

Two Things Worth Knowing Up Front

No outside AI service reads your data. The order files, the customs papers, the sensor readings and the reasoning about them all run on the machines your plan names (your own servers or cloud account, or Runink’s shared machines) — FACE runs on the Runink TIDE platform, which is what makes that a property of the build rather than a setting somebody has to honour. Nothing goes to an outside model provider. That is the kind of answer a security review asks for before it will let a supplier hold its order data.

The queue is where you decide. Every item arrives with its reasoning and the records it rests on, so you can read why it was proposed before you agree to it. What you approve is what gets carried out, and an item you leave alone stays where it is. Which kinds of work are worth putting through the queue in the first place is a question you answer when you set it up — not a value threshold the software polices on your behalf.

Where These Scenarios Stand

None of the scenarios above is a customer result. They are written from what the software is built to do, in the vocabulary of the people who own the problem, and they have not been run against any customer’s data. Nothing here is a case study, and there are no figures in it, because the figures would be ours and the ones that matter are yours.

Runink PULSE, the market-analysis product, and the TIDE platform FACE runs on are covered in their own papers. They are not on this page, and none of the jobs above is a result belonging to either of them.