SAI Groups

Pick in Aisle but not Scan in Main Bank

Detailed view

Retailers across the globe are facing a growing loss prevention challenge driven by the widening gap between product selection in aisles and payment completion at self‑checkout. As self‑service adoption increases, so does the risk of items being picked in the aisle but never scanned at the main checkout or “main bank.”

The Pick in Aisle but not Scan in Main Bank feature of the SAI Group visual AI platform addresses this challenge by using computer vision to correlate customer activity across the store journey—from aisle interaction to checkout behavior. Built on top of existing CCTV infrastructure, the platform detects situations where an item is removed from a shelf but is not subsequently scanned or paid for at checkout, enabling real‑time intervention and post‑incident reporting.

By closing this visibility gap, retailers can reduce shrinkage, protect margins, and preserve a friction‑light customer experience without relying on intrusive controls or additional hardware.

Why Pick in Aisle but not Scan in Main Bank Is Important for Retail Stores

The Aisle‑to‑Checkout Blind Spot

Modern retail loss no longer occurs only at store exits. A significant portion of shrinkage originates earlier—when products are concealed or carried from aisles and later bypassed at self‑checkout, either deliberately or accidentally. Traditional CCTV and point‑of‑sale systems typically operate in silos, making it difficult to connect these two moments into a single incident narrative.

Self‑Checkout Has Changed Risk Dynamics

Self‑checkout lanes have become one of the most vulnerable points in the store. Customers may:

  • • Fail to scan some items
  • • Move items across scanners incorrectly
  • • Complete partial payments
  • • Walk away without finalizing transactions

SAI Group explicitly identifies missed scans and non‑payment at self‑checkout as a core loss vector addressed by its visual AI platform.

Operational Pressure on Store Teams

Store staff are already balancing replenishment, customer assistance, and safety responsibilities. Manually monitoring aisles and checkouts simultaneously is neither scalable nor reliable. Automating detection of “picked but not scanned” scenarios reduces dependence on constant human oversight while improving response consistency.

How does the Pick in Aisle but not Scan in Main Bank feature work?

Continuous Visual Monitoring Using Existing Cameras

The feature is powered by computer vision models layered over existing CCTV feeds, eliminating the need for specialized hardware. The platform continuously observes:

  • •Customer interactions with products in aisles
  • •Object handling behaviors (removal, concealment, placement)
  • •Activity at self‑checkout stations

This approach aligns with SAI Group’s documented architecture of overlaying machine vision AI on existing surveillance systems.

Correlating Aisle Activity with Checkout Events

When an item is picked in an aisle, the system tracks whether a corresponding scan or payment event occurs at the main bank or self‑checkout. If the visual evidence indicates that the item never appears in the checkout flow, the platform flags this as an anomaly rather than immediately assuming malicious intent.

Real‑Time Alerts and Gentle Customer Correction

In supported checkout scenarios, the system can:

  • •Prompt customers discreetly via on‑screen messages to review missed items
  • •Notify store staff through handheld devices or alerts

SAI Group emphasizes self‑correction over confrontation, enabling customers to resolve genuine mistakes without staff intervention, while still giving teams visibility when follow‑up is required.

Evidence Capture and Incident Reporting

When escalation is necessary, the platform automatically extracts and stores legally admissible video evidence, significantly reducing the time required for post‑incident review and reporting. This capability is a documented feature of SAI Group’s visual AI platform.

Benefits of Using the Pick in Aisle but not Scan in Main Bank feature

Reduced Shrinkage Without Friction

By identifying loss events across the full shopping journey, retailers can address shrinkage without adding barriers such as locked cabinets or excessive staff intervention. The result is loss reduction with minimal impact on the customer experience.

Improved Checkout Accuracy

The feature supports higher checkout accuracy by detecting missed scans and non‑payment scenarios in real time. This reduces both intentional theft and genuine scanning errors.

Better Use of Store Staff Time

Automated detection allows staff to focus on customer service and replenishment rather than constant monitoring. Alerts are delivered only when action is genuinely required.

Scalable Across Store Formats

Because the solution leverages existing CCTV infrastructure, it can be deployed across convenience stores, supermarkets, and big‑box formats without major capital expenditure. This scalability is a core design principle of the SAI Group platform.

Stronger Incident Documentation

Automatic extraction of incident footage simplifies internal reviews and external reporting, helping retailers respond faster and more confidently to loss events.

FAQ

Is this feature focused only on theft?

No. The feature is designed to detect anomalies between aisle picks and checkout scans. It supports correction of genuine customer mistakes as well as identification of deliberate non‑payment.

Does it require new cameras or sensors?

No. The visual AI platform operates on top of existing CCTV infrastructure, which is a documented characteristic of SAI Group’s solution.

How are store teams notified?

The platform sends alerts to store staff over connected handheld devices or systems, allowing timely and proportionate responses.

Is customer experience impacted?

The system prioritizes discreet prompts and self‑correction, minimizing friction and avoiding unnecessary staff intervention.

Can evidence be used for reporting?

Yes. Detected incidents are automatically captured and stored as legally admissible video evidence, simplifying reporting and follow‑up.