Exception based reporting in 2026: how retail loss prevention teams reduce shrink faster
Retail shrink is now nearly a $100 billion problem for US merchants (Source: NRF). Over the same stretch, the average number of shoplifting incidents per year rose 18% in 2024 versus 2023, and violence during theft events climbed 17% (Source: NRF). Loss prevention teams cannot review every receipt or every hour of footage, so the smarter move is to look only at what falls outside normal patterns. That is the job of exception based reporting, and in 2026 it delivers the most value when it is paired with video AI.
Key takeaways
- Exception based reporting (EBR) flags the transactions and events that fall outside normal patterns, so loss prevention teams investigate the few anomalies that matter instead of every receipt.
- POS data supplies the triggers, such as refunds, voids, and no-sale drawer opens, while video AI supplies the visual context that confirms whether an exception is fraud or a false alarm.
- Pairing EBR with an AI Security Guard turns hours of manual review into a detect, verify, and resolve workflow measured in seconds.
- All Star Elite cut cash shrink from 6% to 1% and closed investigations more than 50% faster after connecting its POS data to Spot AI.
- Spot AI is camera-agnostic, so retailers reuse the cameras they already own and are typically live in under a week.
What exception based reporting means in retail
Exception based reporting is a data-analysis method that surfaces outliers, the transactions or events that fall outside expected patterns. In a store, those outliers often include high-value refunds, repeated voids, coupon abuse, or a spike in gift-card sales. Rather than combing through every receipt, loss prevention teams review only the exceptions that warrant a closer look. That focus saves hours of manual effort and points attention where the risk is highest.
On its own, EBR tells you that something looked unusual. It does not tell you what actually happened at the counter. That gap is why leading retailers connect EBR to video, so a flagged refund arrives with the matching clip already attached and an investigator sees the event instead of guessing from a line of transaction data.
The reason this matters is scale. A single register can generate thousands of transactions a week, and a district can span dozens of stores. No team can eyeball every one, and pure manual review means most fraud is caught late, if at all. Exception based reporting inverts that problem by surfacing the small set of events that break a rule, then ranking them by risk. Loss prevention leaders get a short, prioritized queue rather than an endless log, which is what makes a lean team effective across a growing store count.
Key terms
- Exception based reporting (EBR): a method that reviews transaction and operational data and flags only the events that fall outside normal patterns.
- Shrink: inventory loss measured as a percentage of sales, covering theft, fraud, and administrative or process error.
- Organized retail crime (ORC): coordinated theft and fraud carried out by groups rather than individual shoplifters.
- POS exception: a point-of-sale event, such as a no-sale drawer open or a manual price override, that a rule flags for review.
The POS and anomaly triggers that start an exception
Every exception begins with a rule. When a transaction or an operational reading crosses a threshold, the system records an exception and, in a connected setup, pulls the matching video. The table below maps the most common triggers to what they can indicate and to the context that video AI adds.
Trigger | What it can indicate | What video AI adds |
|---|---|---|
No-sale drawer open | Cash-handling abuse or sweethearting | A clip of who opened the drawer and why |
High-value or repeated refunds | Refund fraud or fake returns | Footage showing whether a customer and product were present |
Manual price overrides or excessive discounts | Discount abuse | Visual confirmation of the item and the person |
Repeated gift-card sales or activations | Gift-card fraud rings | A timeline of the transactions and the buyer |
Voids after payment | Under-ringing or skimming | Matching video of the counter at that moment |
After-hours access to restricted areas | Internal theft or intrusion | A real-time alert plus recorded entry |
Because rules are configurable, teams tune thresholds to the patterns that matter most in their format. A convenience chain may prioritize fuel-court and gift-card activity, while a department store may weight high-value refunds and fitting-room zones. For deeper coverage of transaction-level integrations, see how Spot AI handles retail POS integration.
How EBR and video AI work together: detect, verify, resolve
The value of exception based reporting compounds when data triggers and video sit in one system. The workflow follows four steps:
- Detect. POS and operational data stream into the EBR engine, which flags any transaction or event that breaks a rule.
- Verify. The connected video AI attaches the matching clip to the exception, so an investigator confirms the event at the counter rather than inferring it from a receipt.
- Resolve. Confirmed incidents move into a case, where video, receipts, and notes sit in one file for loss prevention, HR, or law enforcement.
- Learn. Patterns across stores feed back into the thresholds, so alert rules grow sharper over time and false alarms fall.
This is where an AI Security Guard earns its place. It detects intent in context, not just motion, so it separates a legitimate return from a staged one and routes only the events worth a human minute. Spot AI reports a single-digit false-positive rate, compared with the roughly 50% false-alarm rate common to traditional video systems, based on a leading security operations center's analysis of more than 20 platforms.
False-alarm fatigue is not a minor annoyance. When most alerts turn out to be nothing, associates learn to ignore them, and the real incident slips through. Tightening context, so an alert reflects a genuine anomaly rather than a shadow or a passing shopper, is what keeps a program credible after the first month. The learning loop also gives leadership cross-store visibility, so a pattern that shows up in one region can be caught earlier in another before it spreads.
Start with the exceptions that already have clear rules, such as no-sale drawer opens and high-value refunds, then attach video to each one. Verifying a flagged transaction against the matching clip is what turns hours of manual review into a decision measured in seconds.
Why exception based reporting matters more in 2026
Two forces make EBR essential this year. First, organized retail crime has grown more coordinated. In the latest survey, 67% of retailers reported involvement of transnational ORC groups, and 64% said they report less than half of store-related theft incidents to law enforcement (Source: NRF). Repeat offenders and plate-linked activity are easier to catch when transaction data and video are joined, which is why many teams pair EBR with license plate recognition for loss prevention.
Second, returns have become a major fraud vector. US merchandise returns are projected to reach $849.9 billion in 2025, about 15.8% of retail sales, and 9% of all returns are fraudulent (Source: NRF). Exception rules built around refunds and returns give teams an early signal, and video confirms whether the customer and the merchandise were ever there. Retailers see the same fraud patterns in curbside and buy-online-pickup-in-store flows, which Spot AI covers in its guide to BOPIS and curbside fraud.
The market has already moved in this direction: 85% of retailers now say they use AI to detect or reduce return fraud (Source: NRF). Exception based reporting is how that AI becomes practical on the floor, because it narrows an ocean of transactions down to the handful an investigator should actually see.
What retailers gain: proof from the field
All Star Elite, a sports-apparel chain running 80 stores, connected its existing POS data to Spot AI video intelligence. Cash shrink fell from 6% to 1%, an 83% reduction, and merchandise shrink dropped from a range of 10% to 15% down to roughly 6%. Investigations moved more than 50% faster through centralized case management, and law-enforcement case timelines compressed from two or three months to about one. The same footage even informed product placement, where relocating best-selling jerseys lifted adjacent sales by 5% to 15%. You can read more retail results on the Spot AI customer stories page.
"The ability to formalize our incident reporting, have all our cases on one database, and attach videos to those cases has been a game changer... cameras, case management, and people counting - it's great having that all in one system."
Andrew Gonzalez, Corporate Director of Loss Prevention and Safety, All Star Elite
These outcomes are customer-reported and typical of a mature program rather than guaranteed, but they show the pattern: when exceptions carry their own video evidence, teams close more cases with fewer people. The table below contrasts EBR on its own with EBR joined to video AI.
Capability | EBR alone | EBR plus video AI (AI Security Guard) |
|---|---|---|
Anomaly detection | Flags unusual transactions | Adds behavioral and perimeter events in context |
Investigation | Analyst searches footage manually | Matching clip auto-attached to each exception |
Case building | Transaction data and video live in separate tools | Video, receipts, and notes in one case file |
Response time | Hours to review | Seconds to verify, with real-time alerts |
Coverage | Register and transaction data | Registers, back rooms, entrances, and parking areas |
Considerations and limitations to weigh
No tool removes risk on its own, so plan the rollout carefully. Assess how easily the software will integrate with older POS or networking equipment, since legacy systems can limit which fields feed the exception engine. Set clear alert thresholds to keep notification volume manageable, because too many low-value alerts erode trust in the program. Invest in staff training so associates understand what the system does and why, and confirm the program fits your organization's data-handling policies. A phased pilot at one or two stores usually surfaces workflow adjustments before a wider deployment.
Run a pilot in one or two stores before scaling. Tuning alert thresholds and confirming POS integration early keeps notification volume manageable and builds the staff buy-in that makes an exception based reporting program stick.
Selecting an EBR and video AI solution
When you compare options, prioritize accurate event recognition, flexible data integrations across POS and sensors, scalable storage, clear dashboards, and responsive support. Favor a camera-agnostic platform that works with the cameras your stores already own, since that shortens deployment and avoids unnecessary capital expense. Spot AI is software-led and ONVIF-compatible, keeps full-resolution video on-site in the IVR while only metadata crosses the network, and is NDAA-compliant, SOC 2 certified, and PCI-clean. Most retailers are live in under a week, including configuration, testing, and training. Ask each vendor how it handles multi-store rollouts, how quickly its rules can be tuned, and whether investigators can build a shareable case without leaving the platform, because those details decide whether the program saves time or simply moves the work around.
Ready to see exception based reporting in action? Spot AI joins POS data to video AI so your team verifies exceptions in seconds and closes cases faster, using the cameras your stores already own. Book a demo to see how quickly you can go live.
Frequently asked questions
What is exception based reporting and how does it work in retail?
Exception based reporting scans transaction and operational data and flags the events that fall outside normal patterns, such as suspicious refunds, voids, or no-sale drawer opens. Loss prevention teams then investigate only those anomalies instead of every receipt. When it connects to video AI, each flagged exception arrives with the matching clip, so investigators confirm what happened in seconds.
How quickly can retailers deploy EBR with video AI?
Because Spot AI is camera-agnostic and cloud-connected, it works with the cameras a store already owns, so most retailers are live in under a week. That window usually covers configuration, testing, and user training. A short pilot at one or two locations helps tune alert thresholds before a wider rollout.
What types of exceptions can these systems detect?
Common examples include no-sale drawer opens, excessive discounts or manual price overrides, repeated gift-card sales, high-value refunds, and entry into restricted areas. Operational anomalies, such as unusual inventory movement, can also trigger an exception. Rules are configurable, so teams tune them to the patterns that matter in their stores.
How do I measure the ROI of an exception based reporting program?
Track shrink as a percentage of sales, case-closure speed, recovered merchandise value, and inventory accuracy. Comparing these metrics before and after rollout gives a clear ROI picture. All Star Elite, for example, cut cash shrink from 6% to 1% and closed investigations more than 50% faster.
What challenges should retailers plan for when adopting EBR and video AI?
Common hurdles include integrating with older POS systems, tuning alert volume so teams are not overwhelmed, and earning staff buy-in. Confirming that the program fits your data-handling policies is also important. A phased pilot and hands-on training help address these issues early.
About the author
Sud Bhatija is COO and Co-founder at Spot AI, where he scales operations and GTM strategy to deliver video AI that helps operations, safety, and security teams boost productivity and reduce incidents across industries.
