Time to reclaim the 80% of theft that industry is missing

On the back of Euroshop, the World’s number one retail trade fair, IGD’s Retail Futures Senior Partner, Toby Pickard, sits down with Daniel Gabay, CEO and Co-Founder of Trigo to discuss shrink prevention technologies.

At the trade show, Trigo won the RETA award for Netto Marken-Discount’s success with its Loss Prevention Vision AI solution to reduce theft.

In this exclusive interview, we discuss where in the store, and how Trigo’s technology is helping retailers address the significant challenge of instore theft.

1. What specific types of shrink does your solution reduce, and what measurable results have you achieved in grocery environments?

Most loss prevention technologies focus only on what happens at self-checkout, but our data shows that shrink rarely begins there - it starts in the aisles. That’s why Trigo’s approach begins with monitoring high-risk, highly stolen items at the shelf, observing how customers interact with them and whether they are removed with the expectation they will later be paid for, all while fully preserving shopper privacy.

From there, we follow those items through the store and expect to see them properly accounted for at any checkout option - whether at self-checkout or a staffed lane.

When an item that was taken from the shelf should have been paid for but isn’t - whether due to intent or simple oversight - our system issues a real-time alert, again without identifying or profiling any shopper.

In addition to the alert itself, the system can provide a cropped image or short video clip of the specific action, serving either as:

• a way to resolve honest mistakes on the spot by helping the shopper correct a missed scan, or

• evidence for later investigation

We address shrink across a broad range of scenarios, including:

• Fake scanning or intentional mis-scanning at SCO or manned lanes

• Items held in hand/ left on the counter etc., and never placed on the scanner

• Products left in the trolley or basket

• Concealed items - in jackets, bags, or under other goods - that never become visible at checkout

• Customers who skip checkout entirely, detected through real-time exit-gate monitoring

In grocery environments, this approach consistently reduces losses in top shrink categories - including meat, spirits, OTC medicines, baby formula, and cosmetics - and protects margins where they matter most.

In fact, when analysing top categories such as energy drinks, we found that up to 80% of the verified theft events were not visible to the camera at all, as the drinks were concealed in the aisle before reaching any checkout or exit gate.

2. How does your AI detect suspicious behaviour, and what makes your approach meaningfully different from traditional CCTV analytics or SCO monitoring?

Trigo is different from traditional CCTV analytics or SCO-monitoring in a few ways:

What makes Trigo fundamentally different from traditional CCTV analytics or SCO monitoring is our multi-layered approach:

i. Beyond SCO Visibility

Traditional SCO analytics only see what is visible to the self-checkout camera. As a result, they miss many common loss events - items held in hand, left in the trolley, placed behind other products, or concealed. Trigo overcomes this by monitoring the full shopping journey, not just the POS camera angle.
ii. Behaviour + Product Recognition

Most SCO-focused systems rely on pattern recognition alone - for example, detecting whether an item moved from one side of the scanner to the bagging area.

Trigo adds a second layer: product-level recognition, verifying that every item present at the point of sale is actually paid for - even if it is simply held in the shopper’s hand or resting on the counter without ever entering the scanning field.

iii. High-Risk Item Monitoring (Hidden Items Too)
We apply a third layer of protection by continuously monitoring highly stolen goods from the moment they are picked up. This allows us to detect events even when items are completely hidden from cameras at checkout and would otherwise be invisible to traditional systems.

iv. Effective Even in Kiosk-Style and Unstructured POS Environments

Another advantage of Trigo’s approach is that it works even where scanning at self-checkout is unstructured, such as kiosk-style counters or small-box formats like C-stores, drugstores, and health & beauty stores. In these environments, shoppers often take and return items to the same area, and the shopping flow is less organised. Traditional SCO analytics struggle here because pattern recognition alone breaks down.

With Trigo’s product-level tracking and multi-layered validation, retailers can receive accurate real-time shrink alerts even in these compact, fast-moving store formats.

This multi-layered approach - aisle monitoring, item recognition at checkout, and high-risk product tracking - allows us to detect a far wider range of loss scenarios than traditional solutions. It also delivers great value to security teams analysing shrink offline and providing post-event investigations and trends analysis.

3. What are your accuracy rates in real stores, and how do you minimise false positives to avoid unnecessary associate interventions?

We take accuracy and false positives very seriously. Rather than optimising for a single static accuracy number, Trigo is designed to balance accuracy and alert volume based on each retailer’s policies and risk tolerance.

We use tiered alerting based on confidence and severity. For example, self-checkout alerts can range from a shopper nudge to staff intervention to temporarily blocking checkout. Similar tiering exists for staff-facing alerts. We work closely with retailers to adjust these settings over time based on results and risk appetite.

Accuracy in real stores is influenced by factors such as store layout, traffic patterns, assortment, and operations, which is why retailers can tune sensitivity thresholds to find the right trade-off between detection and unnecessary intervention.

Importantly, the system improves over time as it learns each store’s environment, reducing false positives and increasing confidence without adding friction for shoppers or staff.

4. What operational changes should store teams expect, and how do you ensure alerts don’t overwhelm associates?

Store teams do not have to make any material operational changes. Trigo simply plugs into existing CCTV and POS/SCO systems. Trigo does not require store downtime, or remodelling, or new cameras. Trigo does require a new server or workstation.

Staff will receive real-time alerts on their tablet, mobile device, and desktop. Trigo’s alerts can be used by in-store staff, security, as well as out-of-store security and investigative teams.

After using Trigo, retailers gain insights into how they can improve their loss prevention operations, such as SCO alerts, security staffing, store layout, staff training, and intervention/deterrence policies.

To ensure alerts don’t overwhelm associates, we create a tiered alert system based on confidence and severity. We work with retailers on adjusting the dial based on their results and risk appetite. That way, staff can focus on the most accurate and highest-priority alerts.

5. What infrastructure is required - can your solution use existing cameras and systems, and how does it integrate with POS, SCO, or VMS platforms?

Trigo simply plugs into existing CCTV and POS/SCO systems. Trigo does not require store downtime, or remodelling, or new cameras. Trigo does require a commodity workstation in the store and internet connectivity for minimal cloud processing of certain activities. Trigo integrates with POS systems and SCOs through a secure data API.

6. How do you address customer privacy, regulatory compliance, and responsible use of AI (e.g., no facial recognition, data retention, bias testing)?

Trigo is designed for privacy and is GDPR compliant. Trigo never uses biometric data to uniquely identify individuals, such as facial recognition. Trigo’s video evidence is automatically face-blurred. Trigo does not make use of sensitive biometric data to identify shoppers, nor store it.

7. What ROI can retailers expect, and how quickly do most stores reach payback?

When evaluating payback, we look at the monthly SaaS cost of Trigo together with the one-time workstation investment, amortised over a typical five-year period. We then compare this to the losses the retailer is able to recoup once Trigo is deployed.

While ROI naturally varies by store size and monthly revenue, the pattern we see across grocery environments is very consistent: stores typically reach payback extremely quickly - often within the first 5 -10 days of each month.

From that point forward, the retailer is already net-positive for the remainder of the month, meaning the system effectively generates pure savings.

In many cases, Trigo not only covers its entire monthly cost in the first week but continues to deliver significant incremental loss reduction throughout the period.

This rapid payback is driven by our ability to detect high-value shrink events that had previously gone unnoticed — especially around highly stolen goods, concealment scenarios, and items that remain invisible at checkout.

8. How do you expect AI-driven theft reduction technologies to evolve over the next 3–5 years, and what capabilities do you believe will reshape how grocery retailers prevent loss in the future?

Over the next 3-5 years, the digital twin - a real-time virtual representation of the physical store - will become table stakes, a foundational layer that every retailer operates on.

The real differentiation will come from the physical AI it is integrated with. These models will become more sophisticated at interpreting, predicting, and acting on what's happening in the store, as we're already doing with loss prevention today.

The technology will increasingly move from simply observing the store to directly impacting it by triggering physical action - whether that's responding to theft in real-time, optimising shelf replenishment, or addressing other emerging issues that were previously invisible to store operators.

Looking for more?

Retail Analysis subscribers can access our insight on:

1. The hyper-connected store: the report highlights why the industry must fast-track tech investment to digitalise stores, enabling data-rich, hyper-connected environments that boost efficiency and profit.

Share your story with the food & grocery industry

We welcome contributions from retailers, brands, and solution providers shaping the future of food and grocery. If you have a compelling case study or perspective on technology, operations, or shopper behaviour, we’d be keen to hear from you.

Please get in touch with [email protected]