Transaction Audit
POS + RIQ edge AI · proactive audit

Every till exception, with the video that proves it.

RetaliQ Transaction Audit lines up each POS line with what the RIQ camera above the checkout actually saw. Unscanned items, swapped barcodes, refunds with nobody there, drawers opened for no reason: flagged as they happen, with the clip, the value at risk and the cashier.

Demo · simulated transactions · synthetic camera
Today so far ·
How it works

Two witnesses for every transaction

The POS says what was rung. The camera says what happened. When they disagree, you get an exception with the evidence attached.

1

POS journal

Every line, void, refund, no-sale, override and tender, read live from the POS or its log files.

2

RIQ over the checkout

The edge AI box follows hands, items, the scanner, the drawer and who is on each side of the counter. Video stays in the store.

3

Fusion rules

Both streams are matched second by second against the alert catalogue, with thresholds per client and per store.

4

Alert + clip

Only the exception and its short clip leave the store: to the audit desk, a store manager's phone, e-mail or your case system.

No face recognition: roles come from position and uniform; ages are estimated bands, never identities.

Exception theatre

Watch each exception happen

Synthetic overhead camera of a checkout, driven by the same rules a RIQ runs. Turn the AI layer on and off, scrub the timeline, jump to each alert, export the clip.

0:00
Alert catalogue

Every alert we can raise

From a single unscanned item to a cashier pattern across stores. Each one has its data sources, a default severity and thresholds you tune per client.

Where alerts go

Cashier display

Real-time prompt at the lane (check the cart tray, ask for ID) before the sale closes.

Store manager's phone

Push with the 4-second clip, as in the RIQ Alerts app, for critical cases.

Audit desk

Every exception with clip, status and cashier risk score, across stores and clients.

E-mail, Teams, webhook

Daily digest per store, and an API for case management, ERP or loss-prevention tools.

Audit desk

Last 7 days, ready to review

Click a row for the details, ▶ to watch the clip. Statuses you set stay in this browser.

Exceptions per day

Cashier risk score

Weighted by severity against peers. A score is a reason to look, not a verdict.

WhenStore · laneCashierExceptionItemValueStatusClip
Inventory-driven audit

From the count result straight to the clips

Upload a stock-count result. Every short SKU becomes an automatic search over the period's voids, refunds, unscanned items, substitutions and tag removals at that store. What the tills cannot explain points to the door, receiving or the supplier.

1

Download a demo count

Three stores, one month (12 Oct → 11 Nov 2026). Same simulated data as the audit desk.

2

Upload it, or your own export

CSV or Excel with SKU (or EAN) and book + counted quantities, or the variance. Headers in English, Spanish or Dutch are recognised. The file is read in your browser and never uploaded.

What it needs

Built on what the store already has

POS feed

Live journal (printer or virtual journal tap), transaction log files or the POS vendor's API. Lines, voids, refunds, no-sales, overrides, tenders and the operator.

One camera per lane

Overhead or high-angle view of belt, scanner, drawer and both sides of the counter. Existing Hikvision, Dahua or Sunell cameras usually work.

A RIQ in the store

Edge AI box that reads the cameras and the POS feed, applies the rules and keeps the video local. One RIQ covers several lanes.

Optional

EAS gates and detachers (Amersec), self-checkout logs, scales, RFID tags, door cameras and the stock-count results.

This page is a demonstration: transactions, stores, cashiers and inventory counts are simulated, and the camera is a synthetic rendering. Detection rules and thresholds are illustrative and are set per client during a pilot.