The headline numbers, and what they are actually measuring
Co-op and Devon and Cornwall Police have rolled out SentrySIS, a crime reporting platform, across every Co-op store in the county after a pilot nearly doubled the share of retail crime cases that get solved. Crimes reported through the system are resolved at almost 60%, compared with the force's average solve rate for retail crime reported through traditional channels, and the average time to a completed investigation outcome fell by more than a third, to 35 days.
What makes this worth a retail technology leader's attention is not the solve-rate number by itself, which is a police operations metric more than a retail one. It is that the retailer's own reporting behavior, not a new surveillance camera or detection model, is what moved that number. Co-op changed how it hands evidence to police, and the solve rate followed. That is a workflow result, and workflow results tend to be the ones a retail security team can actually reproduce without a multi-year technology bet.
The mechanism is front-loaded evidence, not smarter detection
Store staff use SentrySIS to upload incident details, CCTV footage, photos, and a written impact statement at the point a crime is reported, rather than filing a bare report and waiting for an officer to come back and request the evidence later. Police can then record and allocate the case in as little as 30 minutes, down from up to seven days under the previous process, because the investigating officer already has what they need the moment the report lands on their desk.
That seven-day figure is the real story here. A week of delay between a reported crime and an officer actually being assigned to it is a week in which footage can be overwritten, memories fade, and a repeat offender commits additional offenses before anyone connects the pattern. Compressing that gap to 30 minutes is not a marginal efficiency gain. It is the difference between evidence that is still useful and evidence that has already degraded past the point of being actionable.
The repeat-offender data makes the case for prioritization, not just speed
Devon and Cornwall Police data tied to the rollout shows that 90% of identified shop theft offenders are repeat offenders, and that just 50 individuals account for roughly a quarter of all shop theft offenses in the force area, with each of those individuals often carrying more than 100 known offenses. SentrySIS supports linking multiple reported offenses to the same person, which means faster individual case processing compounds into something more valuable: an actual intelligence picture of who is driving the bulk of retail crime in a given area.
That concentration ratio is the strongest argument for this kind of platform over point-in-time detection tools. A system that only tells you a theft happened, without connecting it to the same person's prior offenses, treats every incident as independent when the underlying reality is that a small number of repeat offenders are responsible for a disproportionate share of losses. Any retail security investment that cannot support that kind of cross-incident linkage is solving a smaller problem than the one actually driving shrink.
The baseline this is measured against still matters
Co-op reports that overall retail crime is down almost 20% year on year, with anti-social behavior and verbal abuse incidents down 35% and physical assaults falling by a double-digit percentage. Those are genuinely good trend lines, and Co-op's own director of campaigns, public affairs and policy, Paul Gerrard, described them carefully as green shoots with crime levels remaining higher than before the pandemic, rather than claiming the problem is solved.
That framing is worth keeping when evaluating any vendor's results from a pilot like this one. A 20% year-on-year improvement against a baseline that is still roughly double pre-pandemic levels is real progress and also a reminder of how far retail crime has to fall before it reaches a historically normal range. Enterprise buyers should ask vendors for the same two-part framing Co-op used here, the recent trend and the longer baseline, rather than accepting a single favorable percentage in isolation.
What the people closest to the pilot actually said
Devon and Cornwall Police Business Crime Sergeant Chris Chaloner called the pilot results extremely impressive, and a Co-op store manager in Exeter, Alastair Woodburn, said the system has fundamentally changed the way we report crime, a specific and operational claim rather than a marketing line. Policing Minister Sarah Jones also weighed in, framing the collaboration between the retail industry, police and government as exactly what she wants to see, which signals this kind of retailer-police data-sharing platform may get policy tailwind beyond this one force.
The system is now live across every Co-op store in Devon and Cornwall, and discussions with additional police forces are already underway, which is the detail that should matter most to a retail security leader elsewhere in the country. A platform that has moved past a single-store pilot into a countywide rollout, with multiple police forces evaluating adoption, has cleared the bar that most retail security technology never reaches: proof that it works at the scale of an actual jurisdiction, not just a demo store.
The actionable lesson for enterprise retail security teams
The specific lesson here is a narrower and more useful one than buy an AI crime platform. The biggest lever in a loss-prevention stack may well be evidence logistics: how fast and how completely an organization hands evidence to law enforcement once an incident occurs, rather than any new detection or surveillance capability. A process that still requires an officer to come back to the store for footage after the fact is likely losing solve-rate performance to the same seven-day gap Devon and Cornwall Police just closed.
Before evaluating any new AI-driven loss-prevention tool this quarter, it is worth auditing your own evidence-handoff workflow against this benchmark: how long does it take, today, from an incident being flagged to a complete evidence package reaching the relevant police force, and does your system support linking that incident to a suspect's prior offenses. Co-op's result suggests that closing those two gaps alone, independent of any new camera or model, can move a solve rate more than most retail security technology purchases ever will.
