Card feeds first
Ramp, Brex, Amex, and standard card feeds arrive as the primary record. The transaction exists before anyone submits anything, so the report is a confirmation rather than a data-entry exercise.
Platform · operations
The expense report is a form asking an employee to re-enter information that already exists — the card already knows the merchant, the amount, and the date. What is genuinely missing is the receipt image and the business purpose, and those are the only two things anyone should be asked for.
What it does
Ramp, Brex, Amex, and standard card feeds arrive as the primary record. The transaction exists before anyone submits anything, so the report is a confirmation rather than a data-entry exercise.
Photographed receipts extracted and matched to the card transaction on merchant, amount, and date. Unmatched receipts and unreceipted spend are both surfaced rather than one being ignored.
Limits, categories, required receipts above a threshold, and prohibited items checked when the expense is submitted rather than discovered by an approver a fortnight later.
Account, department, location, and project applied from history and from the cardholder’s own pattern, so expenses land in the right cost centre without the employee choosing.
By amount, category, and department, with delegation for absence and escalation when it stalls. It reads your existing matrix rather than introducing a second one.
Out-of-pocket claims routed for payment and card spend settled against the statement, with the difference between the two kept clear rather than blended.
Most expense software still models the expense report as the primary object: an employee creates a report, adds lines, attaches receipts, submits it. The card transaction is then reconciled against it afterwards, which is backwards — the transaction is the fact and the report is a description of it.
Inverting that removes most of the work. The transaction arrives from the card feed, gets coded from the cardholder’s history, and waits for a receipt and a purpose where policy requires one. What the employee does is confirm, not compose.
When policy is enforced by approvers, it is enforced inconsistently — different managers apply different standards, and rejecting a colleague’s lunch is socially expensive enough that most people let it through.
Checking at submission changes who the policy is coming from. The system declines an item outside policy, with the rule stated, before it reaches a person — so nobody has to be the one enforcing it. That is a small design decision with a large effect on whether the policy actually operates.
Most companies track receipt compliance as a percentage and chase the stragglers. The more useful framing is that unreceipted spend above your threshold is an audit exposure and a tax deduction risk, and it accumulates quietly.
Surfacing it as a balance by cardholder, ageing like a receivable, changes behaviour more than a monthly reminder does — largely because it becomes visible to the cardholder’s manager rather than only to finance.
Limits
We read your card programme rather than replacing it. Ramp, Brex, and Amex remain your card provider and your controls sit with them as well as with us.
Booking, itineraries, and travel policy at point of sale belong in Navan or a TMC. We handle the spend once it exists.
Receipt capture depends on somebody photographing the receipt. We make it easy and we surface the gap; we cannot close it for you.
Questions
Tell us your volumes and card programme and we will show you what stops being manual.