Mike Rose, Executive Vice President of Sales – IDIS Americas explores how retailers are using the latest advances in video security and AI analytics.
Retail shrink continues to cost the US industry over $120 billion annually, up from $112 billion in 2022 according to industry estimates.
The average shrink rate has risen to around 1.8% of total sales and retailers now find themselves increasingly challenged by organized retail crime (ORC), aging technology and systemic vulnerabilities.
The Council on Criminal Justice says that between the first half of 2023 and the first half of 2024, shoplifting reports increased by 24%.
Internal theft accounts for around 28% of losses, and partly due to online expansion, returns fraud and fake claims are also on the rise too.
The consequences can be seen in higher consumer prices, demoralized staff and chains shutting down stores due to unsustainable losses.
As profitability comes under pressure, some retailers have responded by reducing labor and opting for self-checkouts and RFID systems.
However, this strategy also brings risks and can increase thefts and error-related losses.
Other countermeasures can impact customer satisfaction: expecting shoppers to wait in line just to enter stores or restricting access to everyday items like washing detergents, toothpaste and baby formulas can negatively affect customer loyalty.
Therefore, strategies are required to safeguard retailers’ staff and inventory while ensuring customer satisfaction.
Video surveillance technology is one of the most valuable tools for combatting these challenges and increasingly, smart video is enhancing retailers’ resilience across every aspect.
AI analytics are letting businesses move beyond passive monitoring, to proactive threat detection and response.
At the same time, they are enabling retailers to drive operational efficiency and even boost staff morale.
AI models cut through noise with object classification and verification and deliver highly accurate alerts and notifications that go beyond eliminating false alarms and missed events.
Intruder and loitering detection, line crossing and other standard security analytics are now considered a baseline requirement when adopting AI video.
Response frameworks can now be based on a broader range of analytics features and functions, with proven, real-time threat detection for unauthorized access, crowd density, loitering, weapon detection, unusual behavior, perimeter monitoring and many more AI analytics functions based on specific risk profiles.
From certain solutions available on the market, retailers can now benefit from multi-layered analysis to verify threats, with cross-references to associated audio and sensor data.
This provides loss prevention teams with the domain awareness they need to initiate faster, informed responses.
Intelligent triggers can activate alarms and automatically send real-time notifications to security personnel, store managers and executives, while initiating protocols and targeting the dispatch of first responders through customizable action plans.
Privacy masking advances that use deep learning are also helping to tackle internal shrink.
This can allow head office staff to view unmasked live footage while keeping images privacy masked on spot monitors and in-store, helping identify internal theft, fraud and ‘sweethearting’ at checkouts.
The adoption of AI technology and AI analytics in surveillance systems is revolutionizing the way retailers manage security across their locations.
A typical US chain store with 500 branches, each equipped with 32 cameras, totals 16,000 cameras to be monitored at any given time.
Traditionally, this would necessitate a vast number of security personnel to keep an eye on every feed, leading to potential oversights and inefficiencies. AI analytics in video serves as a game changer.
By employing highly accurate and sophisticated object classification and behavioral analysis, AI can analyze video feeds in real-time, identifying unusual behaviors or events that human operators might miss.
This enhances the effectiveness of security monitoring while reducing the manpower needed for it.
With AI analytics handling the bulk of the surveillance workload, security personnel can focus on responding to alerts and managing incidents.
Today’s advanced security cameras, such as active deterrent models, are suitable for a wide range of retail environments and have evolved to perform virtual guard tours and auto-tracking, supplementing in-person checks by store security staff.
Grocery and big box stores can also benefit from cameras with built-in red and blue warning lights, microphones and speakers, making it possible to engage with and challenge suspects remotely.
AI analytics and camera performance is underpinned by more user-friendly video management systems (VMS) that make it easier to monitor events in real time; advanced, mobile apps now deliver alerts to loss prevention, security and senior managers on the move allowing them to view live and recorded footage and initiate and direct responses remotely.
Video and AI analytics built into cameras allow enhanced, automated detection of activity including line-cross (single and/or double); loitering; wrong-way motion; left objects and removed objects.
Edge analytics are also now offering functionality such as people counting, face detection and vehicle detection.
These edge AI cameras give retailers flexible and affordable options for upgrading conventional video systems with targeted AI analytics at vulnerable locations providing a phased upgrade migration path that is easier and more affordable.
For loss prevention managers, the availability of more accurate metrics around in-store activity is particularly useful.
For example, the same cameras used to prevent shrinkage can deliver heat mapping to proactively reveal how customers move through stores and around displays, where they spend most time and how behaviors can vary depending on the time of day or day of the week.
This can support more adaptive security strategies, for example with high-risk items being relocated, more closely watched and better protected.
People counting and automated crowd density analysis can help to optimize staff and customer ratios, to get the balance right and ensure that stores are adequately staffed and protected.
These AI analytics can be used in real-time to make operations more responsive to changing customer flows, especially in larger stores where rostering is more complex.
They can also be retrospective, revealing occupancy trends and helping with future resource planning.
While these are things that successful retail operations have always done, AI video is removing guesswork and allows greater consistency with real and actionable data.
Today’s video systems are also providing better value, ensuring faster return on investment.
Measured over the full lifetime of the system, the total cost of ownership (TCO) is influenced by several key factors, including device durability, ease and speed of installation, minimized operating and maintenance costs (such as avoiding license fees and hidden costs often associated with VMS platforms), reduced waste (e.g., by reusing existing infrastructure where possible) and optimized surveillance coverage with the most efficient, application-appropriate cameras.
For example, high performance fisheye cameras that give full scene coverage in ultra-HD can replace three or four fixed lens cameras, enabling savings to be achieved not just from buying fewer cameras, but from faster installation, reduced cabling and infrastructure, and reduced maintenance.
Today’s AI video tech can add even more value by helping to boost profits, not just by protecting them.
The same AI analytics that are providing more information about theft, fraud and retail crime events are also being used to deliver new insights into customer behaviors and activity trends across retail branch networks.
This gives store owners and managers more detailed, more granular and – crucially – more actionable data, while supporting better management decision making.
When it comes to increasing store profitability as well as cutting losses, advanced video systems armed with more powerful AI analytics are now delivering on their promise.
This article was originally published in the June edition of Security Journal Americas. To read your FREE digital edition, click here.