Self-checkout efficiency: balancing cost and customer experience

Retail’s on-off relationship with self-checkouts 

Self-checkout efficiency, how fast, smooth and cost-effective it is to check out without employee assistance and while keeping inventory secure, is always in the news and up for debate. First introduced in 1986, self-checkouts were designed to reduce wait times, increase efficiency and deliver cost savings for retailers.  

For shoppers, the appeal is straightforward: get in, scan, pay, and leave without joining a queue or waiting for a cashier and that is why 54+ of all transactions now go via self-checkout. The challenge for retailers is matching that expectation without sacrificing inventory security. 

The debate is a bit like motorways. We may not enjoy them (congestion, monotony) but without them, journeys take much longer. 

Self-checkout plays a similar role. By the time shoppers place their final item in the basket, they’ve already shifted mentally to what comes next. Their focus is on leaving the store and getting to their next destination—making the checkout feel less like part of the experience, and more like a bottleneck they need to move through quickly. 

Beyond the noise: Why self-checkout isn’t the problem. Scale is.  

Like any new technology rollout, self-checkout has delivered both wins and challenges. Rising shrink is often tied to self-checkout in headlines, frequently citing that 26% of users admit to accidentally not scanning an item, yet this overlooks a critical point: when prompted, 80% of shoppers correct these mistakes. 

For a deeper look at how AI detects and resolves these errors in real time, see our self-checkout security guide. 

Look closer, and the story shifts. Loss at self-checkout hasn’t increased since 2018. What has changed is scale. More shoppers are using self-checkout, and more lanes are in operation driving higher exposure and transaction volume. 

So, are self-checkouts on borrowed time? Or evolving? 

First-generation systems relied on: 

• Barcode scans 
• Weight checks 
• Manual oversight 

They were built for structured interactions. That’s why retailers like Walmart rebalanced their approach, scaling back in some formats while piloting AI-assisted lanes in others. Target introduced a 10-items-or-fewer limit, reporting faster transactions and reduced shrink, with Dollar General and Schnucks following suit. 

At the same time, vision AI proved its value and offered a credible alternative. Delivering measurable ROI by catching missed scans and reducing errors in real time. 

With 80% of grocery transactions still happening in-store, self-checkout isn’t going anywhere. But it is evolving from first-generation systems to intelligent, computer vision-powered experiences. 

Getting the balance right: Efficiency gains powered by people + AI

Self-checkouts offer retailers opportunities to improve both efficiency and customer experience in store. The efficiency gains are well-documented: for operators who get the balance right, self-checkout can meaningfully reduce both transaction time and the staffing cost per customer served. 

As many retailers have learned, success hinges on striking the right balance. See case studies from Intermarche and Edeka. 

The key to achieving the best cost efficiency and customer experience lies in having a workforce augmented rather than replaced by AI. According to SeeChange CEO, Jason Souloglou, “It’s a misunderstanding that AI and other tech means that stores can hire fewer workers. AI frees up employees to conduct higher-value tasks that improve customer satisfaction.” 

This sentiment is echoed by Gerry Hough, of McKinsey, “Retailers are seeing the need for a multipronged strategy. AI is a part of every box, but it is not the same in every box. We are seeing modification and stratification.” 

Lessons from frictionless store experiments 

The journey towards frictionless stores offers a compelling vision of seamless shopping, yet it presents significant challenges in implementation. Amazon’s retreat from just-walk-out technology in 2024 pivoting instead to Dash Carts and a more conventional assisted-checkout model confirmed what the industry had suspected: the technology worked, but the adoption curve was too steep. 

From friction to flow: How computer vision elevates self-checkout 

Computer vision is the use of artificial intelligence (AI) and image processing techniques to analyse visual data from cameras installed, in this context, at self-checkouts. Together with input from other sensors such as barcode scanners and weigh scales, computer vision adds intelligence to the self-checkout process so that errors such as mis scans can be prompted and resolved, while identifying fresh produce on scales can be automated. 

For the shopper whether checking out at a manned or self-checkout the experience should be positive. For retailers looking to adopt computer vision-driven solutions, there are three key considerations: 

Reducing interventions 

Nothing slows a quick self-checkout like a flashing light and a wait for help. Usually it’s a simple issue: a barcode that won’t scan, or an item that misses the bagging area. Instead of blocking the lane and calling a store colleague over, present the shopper instead with a video or image on the screen of the self-checkout to help explain the issue enabling them to fix the problem in a couple of seconds.  

At Intermarché, fixing errors this way cut employee interventions by around 15% with fewer interruptions for shoppers, and a smoother run through the checkout. 

The result is a checkout that feels intuitive. Shoppers keep their pace, queues don’t build up behind a stuck lane, and the usual self-checkout frustration never shows up. Staff benefit too, with fewer call-outs at busy times but the main win is for the customer: getting in and out without slowing their pace. 

Your rules, your way 

Every store is unique, and no store remains the same throughout a day. You need to be able to adapt to the varying demands of your store in real-time according to configurable rules. 

What may be effective during quiet periods may not at peak times. At certain points in the day your aim may be to reduce shrinkage, at others it’s to shorten queues. You need control over this balance to enable dynamic adjustments to overall system sensitivity to shrink so that interventions at self-checkout are in line with your store’s operations and customer needs. 

Intelligence that shoppers love 

Let’s take the example of searching for a ‘fresh item’ at checkout. Typically, it takes a customer between 10–15 seconds to select a single non-barcoded item from the picklist. This leads to 3 in 10 customers avoiding self-checkouts due to perceived complexity and its time-consuming nature. 

In contrast, an AI-powered solution such as fresh produce recognition eliminates the need for manual searches through on-screen catalogues, reducing identification time down to just 0.5s, and the end-to-end time to recognise and add items to the till roll to just 6 seconds. This delivers significant time savings – up to 40% increase in transactions per hour – as employee interventions for struggling customers are minimised. More importantly, 92% of shoppers when given the choice, opt for the new AI-powered approach over manual look-up. 

And why stop at self-checkout? The same recognition system can be deployed on scales in the aisles to deliver a consistent experience for customers and a cohesive, connected approach for retailers. 

The future of self-checkout: smarter, balanced, here to stay 

Self-checkout isn’t going away, it’s being redefined. The next phase isn’t about removing friction entirely, but about managing it intelligently: combining automation with human oversight to deliver speed, accuracy and trust at scale. 

Retailers who succeed will be those who stop viewing self-checkout as a cost lever alone, and instead treat it as a dynamic, experience-led system. One that adapts in real time to store conditions, shopper behaviour and operational priorities. 

The path forward is clear: augment people with AI, reduce unnecessary friction, and design for the way customers actually shop. Get that balance right, and self-checkout moves from being a pain point to a competitive advantage. 

Learn more: 

• For a deeper look at how computer vision addresses self-checkout fraud, loss, and shrink specifically, read our guide to self-checkout security. 

• See how SeeChange’s AI self-checkout software works in practice. Explore the solution or speak to our team.