If your discount stores are getting hit, you already know it's not random
Talk to any LP director running a discount or dollar-store fleet right now and you'll hear some version of the same thing. The same offenders. The same crews. The same product categories. The same handful of locations getting hammered while head office asks why shrink keeps climbing despite tighter controls.
Discount retail has a specific theft profile. Tight margins make every dollar of loss hurt more. Smaller store footprints mean fewer staff, less visibility, and an easier exit. High-traffic, low-supervision environments are where organized retail crime crews like to operate, because they know the math works in their favor. And the violence problem keeps getting worse, which is now a colleague safety issue as much as a financial one.
The LP teams I respect most aren't waiting for the next industry report to tell them what they already see in their stores every week. They're asking a sharper question: what would it take to actually disrupt the people doing this, store by store, in a way that holds up legally and doesn't burn out their people?
The pattern most LP teams see but can't always prove
A small number of individuals drive a disproportionate share of theft and violent incidents. Industry data has been pointing at this for years and frontline LP knows it intuitively. Repeat offenders return to the same stores because they know the staff change, memories fade, and incident records sit in disconnected systems. They lift, they walk out, they come back next week, and the cycle continues
Traditional approaches lean on memory and manual descriptions. That doesn't scale across a multi-hundred-store discount fleet, and it puts unfair pressure on store colleagues who already have too much on their plates. What's needed is a way to recognize repeat offenders the moment they walk in, capture validated incidents with linked video evidence, and share that intelligence between nearby stores in real time. Not a watchlist built on a hunch. A list built on validated incidents, manually reviewed, governed properly, and limited to people involved in confirmed theft or aggression.
What changes when you can see the full pattern
In one major UK convenience and grocery deployment, violent incidents have dropped by more than 90% since SAI's platform went live. That's not a marketing number. That's the difference between a colleague going home worried and a colleague going home. In another UK retailer with a frozen and ambient grocery footprint, shoplifting losses came down by around 50% within four weeks of deployment, largely through the deterrent effect of detection plus efficient evidence sharing. These are real, multi-hundred-store retailers, operating in environments that look a lot like discount stores: high foot traffic, frequent return visits, and a known offender problem.
The capabilities behind those numbers are the ones LP teams have been asking for:
Behavior-based detection at the shelf edge
For concealment, shelf sweeps, and bulk grabs in known high-theft categories.
Revisit alerts
Tied to validated, manually reviewed incidents, with full GDPR governance.
Detection at the exit
For checkout avoidance, gate misuse, and exits through entrances.
Intelligence sharing across stores
So a hit at one location warns the others nearby before the crew gets there.
Violence and aggression detection
Anywhere in the store, with real-time alerts to staff handhelds.
Automated crime reporting
That pulls multi-camera footage into a single video, attaches it to a digital evidence pack, and generates the police statement, instead of leaving your store managers to do it manually at the end of a shift.
The LP director's leverage point
Most discount LP teams are working under a budget that hasn't kept pace with the threat. Adding security guards is expensive and largely reactive. Locking up SKUs frustrates honest customers and pushes legitimate sales away. Visual AI gives you a different kind of leverage: detection and deterrence at the speed of an alert, on top of the cameras you already have.
That last point is worth pausing on. SAI's platform overlays on existing CCTV. No camera refresh. No GPU lock-in. Server deployment in 1 to 2 days, full operation typically inside 4 to 6 weeks. Pilots run in 1 to 3 stores for 8 to 12 weeks, which gives you the data to make the case to your CFO with your own numbers, not someone else's reference deck.
What good evidence looks like to a prosecutor
One detail that LP directors don't always think to ask about: how the system handles evidence. SAI auto-generates a single video from multiple camera feeds, time-stamped, with linked incident records. That gets attached to an MG11-style witness statement automatically. Store colleagues aren't spending the back end of their shift hunting through DVR footage. The case file looks clean when the prosecution team picks it up. More incidents end up actionable because the evidence holds together.
Where to start
If you're running an LP team for a discount fleet and shrink is climbing despite the controls already in place, the next move probably isn't another control. It's visibility into where loss and violence are actually happening, who's coming back, and what evidence you can build on. Start with a 1 to 3 store pilot, pick the locations that are getting hit hardest, and let the data run for 8 weeks. The numbers will tell you whether to scale or walk away.
Most retailers don't walk away.
Lead magnet ideas
The repeat offender playbook: how discount LP teams disrupt the 5% of people driving most of their loss
A 6 to 8 page tactical guide focused on identifying, tracking, and acting on repeat offenders across a multi-store fleet. Covers the role of validated incident data, GDPR-safe revisit alerts, cross-store intelligence sharing, and how to build cases that prosecutors will actually pursue. Includes a self-assessment checklist for current-state LP capability. Lets the LP director hand it to their team and to legal as a single document.
Discount retail's violence problem: a colleague safety briefing for LP and HR
Co-branded asset for LP and HR. Frames the violence-against-staff problem with the 90%+ reduction stat from a major convenience and grocery deployment as the anchor, then walks through what aggression detection, real-time staff alerts, and revisit alerts look like in practice. This is the asset that justifies budget at the C-suite level because it links shrink to retention and to corporate duty of care. Higher conversion potential because it appeals beyond LP.