What happens to a line the policy does not cover?

It is held for a named person with the rule it could not apply stated, rather than passed or failed on a guess, and a missing receipt or field is reported as missing.

Is this the same as supplier invoice intake?

No. That page owns the supplier payable and its approval path, matching an invoice to a purchase order and a receiving record. This page is an employee submission checked against your reimbursement policy, line by line, and it approves, reimburses and pays nothing.

Expense report processing and policy checks

An approver reads a whole report to find the one line that breaks policy

Today

The policy lives in a document nobody opens during the check

An approver reads a whole report to find the one line that breaks policy.
The policy itself lives in a document nobody opens during the check.
A submission comes back twice, because the missing receipt was noticed late.
And nobody can see the state of the queue without asking the person holding it.

With agents

The failing line and the rule it failed arriving together, with an absence named at arrival

The approver decides instead of searching.
A missing receipt is named at arrival.
The policy is applied the same way twice.

How an agent runs it

The reading moves into a check that runs per line against the policy you actually wrote down, and the approver gets the failing line with the rule beside it.

Takes

Receives each submission from the channels you authorize, with the submitter and the date it arrived recorded against it.

Reads

Takes the fields and the receipts attached, and names what is missing rather than inferring it.

Tests

Checks each line against the policy you wrote down, rule by rule, and against only those rules.

Names

States the rule that failed and the line that failed it, so the approver reads a reason rather than a verdict on the whole report.

Assigns

Codes to the account and cost center from rules that are written down rather than remembered.

Holds

Keeps anything the policy cannot decide for a named person instead of guessing, and says which rule it could not apply.

Human gate

Routes every exception, and everything above the threshold you set, to the named approver with the failing line shown. It approves nothing, reimburses nothing, pays nothing and grants no policy exception, and nothing it states is a tax, employment or regulatory determination. The policy and the submissions are yours, so the check can be run by hand at any point without the agent.

Results

What changes when the policy is applied the same way twice

The approver decides instead of searching

The failing line and the rule it failed arrive together.

A missing receipt is named at arrival

The required fields come from your policy, so an absence is a result rather than a later surprise.

The same rules every time

The check runs against the written rules rather than against whoever is reviewing.

Where this fits

Workflow / Automate a Workflow

The Workflow tier, Tier 1, through the Automate a Workflow door: one process end to end, from the submission to the approval a person gives. A workflow proves it, a role scales it, a department transforms it, and an agent-run finance back office is the direction rather than a promised state. How It Works describes the gates and the thresholds.

Tier 1 / Prove

How It Works

Questions, answered

No. It names the failing line and the rule and routes the decision, and approval and anything above your threshold are both gated. The policy and the submissions are yours, so the check can be run by hand at any point without the agent.

AI Opportunity Audit

Bring last cycle's submissions and the policy somebody checks them against, and see how many rules are already written as rules

Bring last cycle's expense submissions and the policy document somebody checks them against, and see how many of its rules are already written as rules.

Book an audit