How AI Helps Retailers Manage Self-Checkout Accuracy and Loss


Self-checkout systems have evolved rapidly over the past decade. Early systems did not have the sophisticated intelligence that today’s self-checkout technology offers. Retailers who are using self-checkout systems or are considering moving to customer self-checkout convenience should investigate the available solutions carefully.

Subramanian Kunchithapatham
Subramanian Kunchithapatham

To better understand today’s self-checkout technology, LPM asked Subramanian Kunchithapatham, vice president of engineering at Sensormatic Solutions, to explain how artificial intelligence (AI) helps retailers provide a better customer experience while helping prevent loss. To read a more in-depth discussion, download the LPM whitepaper “Navigating a Profitable Path on the Journey to a Cashierless Future.”

How have self-checkout systems incorporated artificial intelligence and what are the benefits of AI?

Artificial intelligence has become a critical component of loss prevention in self-checkout technology. Incorporated into the video and scanning systems, AI allows retailers to collect insights into the self-checkout process to predict patterns, identify frictions in checkout, and better understand how theft occurs. AI systems in self-checkout kiosks can detect anomalies in a customer checkout, flag suspicious transactions, and alert staff to suspected shoplifting in progress.

- Sponsors -

AI has many forms. Computer vision and machine learning technologies can be leveraged within self-checkout systems to help prevent loss events. AI can also be integrated within a broader retail technology platform, allowing it to use multiple data sets to inform its predictions and insights. Pulling data from areas like inventory intelligence and traffic measurement, as well as self-checkouts and other points of sale, creates a holistic view of shopper behavior and thus greater shrink visibility.

How can self-checkout systems detect and prevent “item swapping” theft?

Self-checkout systems with intelligent image scanning can predict the bar code of the item in the frame. If a customer passes an item through the frame and the predicted bar code is not processed, but another item’s bar code is scanned repeatedly, the system can alert staff to intervene.

No matter what shopping service they’re utilizing, customer service remains a high priority for customers, as do convenience, efficiency, and well-stocked stores. Intelligent image scanning can help power this, as well. For example, when purchasing non-barcoded items like fresh produce, the system offers a short list of suggestions instead of requiring customers to scroll through a full list of fruits and vegetables on the interface.

What is the benefit of the video display on a self-checkout monitor?

Video displays that show the live feed of a customer going through a self-checkout act as an eye-level deterrent. Generally, customers see the barrier to theft as lower at self-checkouts because they feel like their actions are unseen. When a customer knows that a camera is monitoring their checkout process, that barrier to theft becomes much higher.

How can retailers use audits of transactions to reduce errors?

Retailers can use audits of select transactions to understand the types of items that are not being scanned, parts of the process where customers are able to skip a step and other areas that allow for theft. Regularly collecting data allows AI to better learn which transactions need to be flagged to staff.

Self-checkout audits can also be used to inform overall inventory intelligence, improving the customer experience by giving retailers real-time data about the items they have in stock. Additionally, having the right merchandise available at the right time allows retailers to improve their sustainability outcomes. By reducing overstocks, retailers can avoid unnecessary truck trips, as well as decreasing excess CO2 emissions when shoppers must make multiple trips for out-of-stock items.

What’s next? How does AI go beyond alerting retailers to a theft in progress?

The goal of AI and machine learning loss prevention solutions is to move from identifying a case (either during or after) to preventing a case before it ever happens. When AI data intelligence is integrated with connected solutions and third-party data, as it is through our Sensormatic IQ platform, it can be leveraged into prescriptive actions to improve store performance.

The people aspect of these solutions cannot be overlooked. AI does not replace the staff necessary to manage successful self-checkout systems but rather supports retail employees in better understanding the risks and implementing the right technologies to manage shrink and improve the customer experience. Retailers should be investing in the proper training to help employees maximize the benefits of AI.

Stay Updated

Get critical information for loss prevention professionals, security and retail management delivered right to your inbox.