Strengthening ASCONE with Vision AI for physical retail security

How Vision AI enhances human judgment and operational efficiency

ASCONE provides retailers with a structured, fair, and evidence-based approach to managing suspected theft. Its strength lies in clear observation at every stage — from product selection through to exit. But in fast-paced, high-traffic store environments, maintaining this continuous visibility relies heavily on sustained human attention.

This is where technology can strengthen the process. New advances in computer vision allow retailers to support ASCONE from the earliest stages, helping teams maintain clarity and consistency while easing operational load. Instead of replacing human judgment, these tools ensure that security professionals can focus their expertise where it matters most.

Human-led ASCONE is effective but difficult to maintain

ASCONE is simple in theory but demanding in practice. Each step is critical to ensure that when a suspected shoplifter is approached, the evidence is clear and defensible:

• Approach – see the customer as they move toward a product
• Selection – watch the moment an item is removed from the shelf
• Concealment – observing a product being hidden
• Observation – maintaining uninterrupted surveillance of the suspect
• Non-payment – confirming they bypass the checkouts without paying
• Exit – acting only once the individual has left all points of sale

In retail training, it is repeatedly stressed: losing observation of the individual at any stage means the entire ASCONE chain is broken. A stop cannot be made, no matter how suspicious the behaviour. This is good – it protects both the customers’ rights and the retailer’s liability.

Yet in real life, observation is incredibly hard to maintain. Stores have multiple entrances and exits, crowded aisles, blind spots, and staff spread thin. For Security Operations Centre (SOC) operators, the challenge multiplies across tens or hundreds of stores.

By applying advanced computer vision technology to the early stages of ASCONE — Approach, Selection, Concealment (ASC) — we can enable Security Operations Centres (SOCs) to focus on the latter half of the process: Observation, Non-payment, Exit (ONE). The result is a partnership between vision AI and human decision-making that enhances accuracy, eliminates bias, and scales security efficiency across entire estates.

The typical SOC operator spends six of their eight working hours trying to spot Approach, Selection, and Concealment. Only around one hour per shift is spent monitoring the later stages where their human judgment is most valuable. (The final hour is usually taken up by administrative tasks).

This imbalance highlights a critical issue: well trained, highly skilled humans excel at making decisions in complex scenarios, but are less effective at the constant, repetitive scanning required to spot early indicators of theft.

Empowered with Vision AI: Surfacing the ASC

By using computer vision to recognise Approach, Selection, and Concealment, Vision AI systems provide real-time alerts directly to SOC operators.

This is how it works:

• Automated detection: When Vision AI detects a concealment event — for example, a product placed in a bag, under clothing, or hidden in a trolley — the AI system creates an instant alert with a clear, timestamped evidence video
• Focused monitoring: Instead of scanning dozens of screens for incidents, operators can be immediately pointed to the relevant camera feed, and the adjacent streams which allow them to maintain Observation
• Evidence in hand: The system provides clear evidence of Approach, Selection, and Concealment, allowing operators to confidently shift focus to the ‘ONE’ steps of the standard.

Rather than spending their day scanning endless CCTV feeds, SOC operators can now dedicate their time to doing what they do best: maintaining observation and making timely interventions.

Vision AI systems put the ‘ASC’ on autopilot — so humans can concentrate on the ‘ONE’.

The benefits of vision AI-enhanced ASCONE

Augmenting ASCONE with Vision AI delivers a wide range of operational benefits for retailers and security teams:

1. Unbiased monitoring: Human observation is prone to bias. Subconscious judgments about age, gender, clothing, or ethnicity can influence who is monitored more closely. Vision AI systems can reduce this risk by focusing entirely on behaviours, not individuals. A concealment is flagged regardless of who performs it, ensuring fairness and compliance with equality standards.
2. Real-time alerts: Alerts are delivered as soon as a concealment is detected, which means retailers can prevent losses before they reach the store exit. This is reactive, real-time, loss prevention.
3. Scalable insights: Beyond real-time monitoring, Vision AI provides aggregated data on concealment patterns. Retailers can answer critical questions:
   1. Which stores experience the highest concealment rates?
   2. At what times of day or week does it happen most?
   3. Which product categories are most targeted?
      This is proactive, data driven, estate-wide loss prevention. These insights allow for smarter deployment of resources, aligning staff coverage with actual shrink risks.
4. Scalable efficiency: With Vision AI systems detecting ‘ASC’, existing security operations setups can cover far more locations without additional staff. For retailers managing large estates, this means improved security without inflating costs. For SOC operators, it means doing more with the resources you already have.
5. Reduced manual effort: Instead of arbitrarily monitoring camera feeds, operators receive curated alerts. This allows them to focus their attention where it matters — on observing suspects through the ‘ONE’ stages and ensuring interventions are precise and effective.

Why vision AI enhanced ASCONE matters for retailers

Retail shrink costs UK retailers billions per year. While ASCONE provides a robust framework for lawful intervention, its success hinges on human observation — where no operator, however skilled, can guarantee perfection.

By enhancing ASCONE with Vision AI, retailers unlock a new model of security: one that is consistent, scalable, and resistant to human limitations. Importantly, it doesn’t replace human expertise – but augments it. Systems such as SeeChange’s Connected Store platform ensure security operators are equipped with the evidence they need, precisely when they need it.

This human-AI collaboration doesn’t just reduce losses. It also boosts staff morale. Security officers no longer need to spend hours in tedious scanning mode but can instead engage their judgment and training in decisive moments where they add real value.