Top 10 AI Prompts and Use Cases and in the Retail Industry in San Bernardino

San Bernardino retailers sit at the intersection of rising customer expectations and tighter margins, and AI prompts are the practical bridge between data and action: a well-crafted prompt can turn inventory feeds, point-of-sale signals, and camera streams into real‑time pricing, personalized offers, fraud alerts, or loss‑prevention cues.

Local businesses in the Inland Empire are already leaning on technology more than ever, so prompts that drive AI shopping assistants, dynamic pricing, and visual-search workflows - outlined in Insider's roundup of the top AI retail trends - deliver measurable gains; at the same time, AI‑powered surveillance and real‑time threat detection help keep stores safe without spoiling the shopping experience, as Trackforce explores in its 2025 security trends.

For managers and staff who need to write prompts that work, Nucamp's AI Essentials for Work syllabus provides hands‑on training and a practical syllabus to apply these tools on the shop floor and online (see the AI Essentials for Work syllabus for details).

Attribute	Information
Description	Gain practical AI skills for any workplace. Learn how to use AI tools, write effective prompts, and apply AI across key business functions, no technical background needed.
Length	15 Weeks
Courses included	AI at Work: Foundations; Writing AI Prompts; Job Based Practical AI Skills
Cost	$3,582 early bird; $3,942 regular. Paid in 18 monthly payments, first payment due at registration.
Syllabus	AI Essentials for Work syllabus - Nucamp
Registration	Register for the AI Essentials for Work bootcamp at Nucamp

Selection of the top 10 prompts began with practical impact for California retailers - prioritizing use cases that address rising shrink and safety risks, integrate with existing camera and POS systems, and meet new state compliance expectations - so projects that clearly reduce losses or speed response ranked highest.

Primary criteria included demonstrated loss‑prevention benefit (Coram's guide documents a 26% jump in shoplifting and notes retailers are increasing tech investments), real‑time operational effect (camera systems that

send real‑time alerts instantly

make weapon or plate detection actionable), legal fit under California's evolving ADMT/CCPA rules (pre‑use notices, opt‑out rights, and the January 1, 2027 timelines for ADMT disclosures must be respected), the need for documented risk assessments and phased cybersecurity audits starting in 2028, and vendor interoperability for multi‑site rollouts common in the Inland Empire.

Each candidate prompt was scored on measurable ROI, privacy and auditability (favoring privacy‑by‑design workflows called for by regulators), ease of staff adoption, and scalability across store footprints; lower‑scoring items were kept as future pilots.

The result: a prioritized list that balances loss prevention, customer experience, and regulatory safety so San Bernardino retailers can deploy high‑value prompts without creating new compliance headaches - turning surveillance and inventory data into seconds‑counting alerts that protect people and profits.

For more on the loss‑prevention playbook, see the Cybersecurity Fundamentals registration page for related guidance, the AI Essentials for Work syllabus for ADMT/CCPA compliance context, and the AI Essentials registration page for privacy‑first tooling and best practices.

For San Bernardino retailers looking to close the gap between detection and response, Coram's real‑time weapon detection prompt turns existing IP cameras into an early warning system that delivers confirmed alerts in under five seconds, uses multiple AI models to keep false positives low, and even lets staff “track the journey” of a suspect across cameras when a face isn't visible - so teams can act on seconds, not minutes.

Designed to work with Coram Point and set up in minutes without extra hardware, the system runs proprietary foundation models locally (with cloud review to reduce errors) and offers per‑camera, month‑to‑month licensing so pilots can start small and scale.

For retailers balancing safety, budget, and California compliance timelines, this approach reduces dependence on third‑party monitoring, speeds frontline response, and integrates with familiar camera feeds; see Coram's weapon detection overview for feature details and the Coram company profile for background on the vendor's Sunnyvale roots and patents.

Attribute	Detail
Alert speed	Under 5 seconds (confirmed alerts)
Camera compatibility	Works with existing IP cameras
Deployment	Coram Point; no extra hardware required
Licensing	Per‑camera, month‑to‑month; start with one camera
Tracking	Journey tool tracks suspects across multiple cameras
Company	Founded 2022; HQ: Sunnyvale, CA

License Plate Recognition (LPR) prompts - the kind retailers can run alongside existing camera‑based loss‑prevention systems - offer a practical route to turn vehicle sightings into actionable leads: when an LPR cue flags a suspicious plate near a shoplifting incident, staff and local law enforcement gain a focused search token rather than sifting through hours of footage, which makes recovery and follow‑up more efficient.

In San Bernardino, where stores are already deploying computer‑vision tools to curb shrinkage, pairing LPR prompts with those workflows is a natural extension of proven loss‑prevention tactics (computer vision loss prevention in San Bernardino retail).

Small chains and independent grocers can also offset pilot costs by tapping local funding programs and workshops that underwrite AI trials, making an LPR rollout more affordable and easier to evaluate before wider deployment (San Bernardino AI grants and workshops for retail pilots).

The result: a focused, privacy‑minded tool to narrow searches and protect margins without overwhelming staff with noise.

Irisity's IRIS+ platform equips San Bernardino retailers with behavior‑aware prompts that turn existing IP cameras into active “AI agents” - flagging anomalies, loitering, line‑crossing, unattended objects, crowd density shifts, and more in real time so staff see only verified events instead of hours of noise.

Its adaptive anomaly detection learns scene norms and surfaces unusual patterns without manual rule‑writing, while privacy‑first tools such as pixelation and dynamic anonymization help meet modern data‑protection expectations; operators get fast forensic search, BI dashboards, and per‑camera analytics that scale from a single store to multi‑site rollouts.

Proofs‑of‑concept deploy in minutes with the IRIS+ Simulator and a 30‑day trial, and real‑world deployments (from transport hubs to retail) demonstrate both safety and operational gains - for example, Irisity's systems helped stop 29 suicide attempts on the Stockholm metro - making the platform a practical choice for retailers who need high accuracy, lower false alarms, and audit‑friendly telemetry.

Learn more about Irisity's AI Video Analytics and its Anomaly Detection capabilities for retailers.

Feature	Detail
Anomaly Detection	Real‑time, scene‑adaptive alerts with no manual definitions
Anonymization & Privacy	Black/dynamic/background masking and pixelation (GDPR‑grade)
Deployment	On‑prem, cloud, hybrid or air‑gapped; edge support
Camera Compatibility	Works with any IP camera (RTSP/ONVIF), extends legacy systems
Proof of Concept	IRIS+ Simulator: deploy in ~20 minutes; 30‑day trial available

“Irisity is a video analytics company that develops AI-based solutions for security cameras. Our software IRIS can detect security threats such as intrusions, violent behavior, falls, flames and loitering. IRIS protects integrity with our patented anonymization technique that ensures accurate detections without exposing any personal information.”

Rhombus' audio analytics prompt - centered on the A100 Audio Gateway - adds a practical layer of sound‑based intelligence for San Bernardino retailers by turning noise into timely, actionable alerts: gunshots, glass breaks, shouting or aggressive speech, and other abnormal audio are flagged in real time and pinned to synced camera footage so managers can jump straight to the exact second an event occurred; the gateway can also broadcast prerecorded messages or enable two‑way intercom to de‑escalate incidents or announce evacuations from the console or mobile app.

Built‑in AI indexing and policies (loud sound, tamper, glass break beta) reduce false alarms, onboard storage holds up to 50 days with optional cloud archiving, and the A100 can act as a standalone sensor hub that integrates with cameras and Rhombus' growing operational analytics (Line Crossing, Occupancy Counting) to turn safety signals into staffing and flow insights.

For stores balancing privacy, response speed, and ease of rollout, Rhombus' docs explain console controls and licensing for audio features, while the product announcement lays out the detection and operational use cases in detail.

Attribute	Detail
Device	Rhombus A100 Audio Gateway product announcement and overview
Detections	Gunshots, glass break, shouting/distress, abnormal noise levels, keywords
Features	Two‑way audio, prerecorded broadcasts, AI audio analytics, sensor gateway
Storage & Compliance	Onboard audio up to 50 days; optional cloud; 10‑year warranty; FCC/NDAA/TAA listed
Policies & Controls	Console toggles for listening/recording, loudness policies, tamper alerts (Rhombus Audio Gateway walkthrough and support guide)

“The A100 is a truly ground‑breaking audio device that captures clearer audio, provides two‑way audio support, and layers AI‑based audio analytics, so users can be notified when a gunshot is detected, or a fight breaks out. When you're dealing with safety, seconds matter and the A100 provides proactive data and features to help defuse security issues in real‑time.” - Omar Khan, co‑founder and CTO of Rhombus Systems

Genetec's Security Center Omnicast turns a bank of camera feeds into a searchable, multi‑camera investigation engine that's especially useful for multi‑site San Bernardino retailers who need to trace an incident across a footprint in seconds - visual tracking follows people or moving objects across cameras, while forensic tools let operators search by bookmark, motion, or date and time and even compress hours into a short Video Synopsis for rapid review.

Built as an IP‑based VMS with on‑prem, hybrid, or cloud‑extended storage options, Omnicast supports chain‑of‑custody evidence export, ALPR and transaction integrations, and built‑in privacy tools so loss‑prevention teams get usable leads without wading through noise; see the Genetec Security Center Omnicast product overview and watch the Genetec Omnicast demo and product videos for examples of the workflow in action.

Attribute	Detail
Key capability	Genetec Security Center Omnicast visual tracking and multi-camera search
Search modes	Bookmark, motion, date/time, forensic search
Deployment	On‑prem, hybrid, or cloud‑extended storage
Operational help	Genetec Omnicast demo and product videos

For a retailer, that means turning a chaotic ten‑minute incident into a pinpointed clip and a ticketable lead - a literal time‑saver when every minute matters on the shop floor.

Gorilla Technology's person‑tracking prompt turns a network of store cameras into a single investigative thread that's especially useful for California retailers needing to trace suspects across aisles, parking lots and multiple sites: the platform can analyze hundreds of cameras in real time to stitch sightings into a seamless video trail, so a shopper who darts through three departments becomes one continuous breadcrumb for loss‑prevention teams rather than three disconnected clips.

Built around edge AI appliances like IVAR and the EVMS dashboard, Gorilla's stack pairs real‑time facial and license‑plate detection with behavior analytics (loitering, line crossing, crowd flow) and big‑data indexing so searches are fast, auditable, and interoperable with police or VMS workflows.

For San Bernardino chains that must balance fast response, privacy controls, and existing hardware, Gorilla's approach offers scalable cross‑camera tracking without ripping out an entire system - see the Gorilla Video Analytics capabilities, the IVAR recorder deployment overview, and a concise cross‑camera tracking explainer for how the technology links feeds into one searchable trail.

locates persons of interest across multiple geographical locations

See detailed resources: Gorilla Video Analytics capabilities for retail loss prevention, IVAR recorder deployment and overview, and cross-camera tracking explainer for video analytics.

Attribute	Detail
Core capability	Locate persons of interest across multiple geographical locations; analyzes hundreds of cameras in real time
Key features	Real‑time facial detection, LPR, behavior analytics (loitering, line crossing, crowd detection)
Deployment	Edge AI (IVAR), EVMS dashboard, interoperable with existing CCTV/VMS

Eagle Eye Networks' crowd counting and people‑flow prompts turn ordinary camera feeds into business intelligence for California retailers - helpful in San Bernardino when staffing and store layout must flex with unpredictable foot traffic - by offering cloud‑delivered analytics like People Counting, Line Crossing, Intrusion Detection, Loitering and Camera Tampering that can be turned on per camera and viewed from the dashboard or mobile app; best practices (re‑aim a camera to a top‑down view, place the counting line near the center of the image, and set object width/height so people - not carts or displays - are tracked) keep accuracy high, while alerts and historical graphs let managers spot a midday surge or a slow afternoon and respond with schedules or promotions rather than guesswork.

These analytics run without ripping out hardware, integrate with Eagle Eye's cloud VMS and Bridges for larger sites, and export counting graphs and event history so audits and ROI checks stay simple - see the Eagle Eye Networks object counting guidance and the Eagle Eye Networks real‑time analytics app note for setup tips and configuration details.

Capability / Best Practice	Detail
Core analytics	People Counting, Line Crossing, Intrusion, Loitering, Camera Tampering (per Eagle Eye app notes)
Camera placement	Top‑down view preferred; keep persons same size across image for best counting accuracy
Line placement & rules	Place line near center; object must be tracked prior to crossing and ≥50% must cross; only one line supported
Configuration tips	Set object width/height ranges to target people; review green analytics boxes in preview and history
Operational notes	Counting graphs accessible in dashboard; daily count resets at 2:00 AM camera timezone

IntelliVision's high‑accuracy AI video analytics - highlighted on its monitoring‑station and smart‑city pages with a claimed ~99% detection accuracy on public databases - can form the backbone of a fall/duress detection prompt that turns passive cameras into active safety sentries: when the system spots a sudden collapse or prolonged immobility it can send an instant, location‑pinned alert to designated staff so help arrives in seconds rather than minutes (see Oxania's primer on real‑time fall detection).

Paired with automated escalation paths - two‑way audio, prerecorded announcements, or live intervention - stores get a practical, privacy‑minded safety workflow that both reduces medical risk and limits fraudulent “slip and fall” exposures; vendors emphasize privacy controls that avoid facial identification while still delivering actionable clips (Abto's fall module documents real‑time alerts with non‑identifying analytics).

For retailers who need faster human response, integration options range from on‑site staff notifications to services that couple AI detection with live monitoring and rapid police contact, proving the difference between a long lie and a quick rescue when seconds count.

Attribute	Detail
AI analytics reference	IntelliVision AI video analytics for monitoring stations - ~99% accuracy on public databases
Real‑time alerting	Oxania real-time fall detection for retail, hospitals, and warehouses - instant alerts to designated personnel
Fall detection module	Abto fall‑detection example - ~85% accident detection accuracy; real‑time alerts; GDPR‑style non‑identifying operation
Live response option	Deep Sentinel retail live monitoring and guard intervention - rapid response workflows with sub‑30s police contact in reported cases

Digital Watchdog's DW Spectrum turns everyday IP cameras into a fast‑loading, smart‑search workhorse for San Bernardino retailers by using dual streaming and adaptive scaling so staff get crisp clips when it matters and lightweight feeds the rest of the time; the client auto‑switches to low‑res under high CPU or limited bandwidth, but

“snaps” to high‑resolution whenever an item is larger than 172 pixels or a tile is full‑screen - so a manager double‑clicking a suspect clip sees detail, not a blur.

Motion detection is performed on the low‑resolution stream (up to a 1024×768 threshold), and the Motion + Lo‑Res recording mode archives high‑res only when motion occurs, which reduces archive costs while keeping forensic quality for incidents.

Practical admin tips in the guide - add cameras one at a time, set per‑layout resolution, or force motion detection to a specific stream when cameras report unusual profiles - help avoid CPU overload on busy store systems.

For retailers pairing computer vision loss‑prevention tools with efficient VMS performance, see Digital Watchdog's dual‑streaming documentation and a local overview of computer vision for loss prevention in San Bernardino.

Feature	Detail
Dual streaming	Primary high‑res + secondary low‑res for adaptive scaling
Motion detection	Performed on low‑res stream (up to 1024×768)
Recording mode	Motion + Lo‑Res: high‑res on motion, low‑res otherwise
Client auto‑switch	Switches to low‑res for bandwidth/CPU limits or >16 tiles; full screen/≥172px forces high‑res

Milestone XProtect's edge-and-cloud hybrid prompt gives San Bernardino retailers a practical pathway to modernize without ripping out working cameras: run XProtect on‑prem for low‑latency recording and local control, then extend select sites or long‑term archives into the cloud so multisite chains get centralized search, remote access, and elastic analytics on demand - an approach shown in the Arcules–Milestone hybrid offering that makes cloud-connected cameras visible inside XProtect for unified investigation (Arcules Milestone hybrid VMS solution overview).

For retailers who need edge performance and scalable retention, certified edge architectures such as Cisco HyperFlex hosting XProtect with Cohesity archiving explain how to deploy recording servers, preserve chain‑of‑custody exports, and move older footage to S3 for long‑term retention while keeping real‑time analytics at the store level (Cisco HyperFlex Milestone video surveillance edge guidance).

The hybrid model balances capex and ops - keeping high‑priority video local for instant review while offloading bulk storage and cloud analytics - so loss‑prevention teams get faster, auditable leads without ballooning on‑site racks or bandwidth bills.

Attribute	Detail
Deployment options	On‑prem XProtect, cloud/VSaaS integration (Arcules), or hybrid edge+cloud
Key benefits	Low latency at the edge, centralized multi‑site search, scalable cloud archiving, AI analytics
Common integrations	Arcules VSaaS, BriefCam/analytics partners, Cisco HyperFlex edge, Cohesity archiving
Typical on‑site components	XProtect Management & Recording Servers, Smart Client, XProtect Smart Wall

“Milestone resellers and customers can deploy to leverage cloud benefits” - Bjørn Skou Eilertsen, CTO, Milestone Systems

San Bernardino retailers ready to act should start small, pilot fast, and layer expertise: run a focused AI prompt pilot that targets shrink or safety (so a few cameras and one store produce usable, seconds‑counting alerts), pair that rollout with a documented privacy and cyber plan, and bring local talent into the loop - CSUSB's Center for Cyber and AI offers CAE‑level expertise and community workshops to help retailers assess risk and upskill staff, while federal and state funding avenues (recently reinforced by Rep.

Aguilar's $300,000 award to CSUSB) can underwrite trials and curriculum development; for hands‑on prompt skills, the Nucamp AI Essentials for Work bootcamp (15 weeks) teaches prompt writing, tool use, and operational workflows so managers and frontline staff can deploy safe, auditable prompts without a technical degree.

Use local security partners to blend AI with trained guards and schedule a short POC, document ROI, then scale - turning camera and POS data into clear leads instead of noise and keeping compliance and response times top of mind.

Resource	Why it helps	Link / Contact
CSUSB Center for Cyber and AI	CAE‑level cyber research, workforce workshops, local expertise	CSUSB Center for Cyber and AI - Cybersecurity and AI programs • cyber@csusb.edu • (909) 537-7535
Nucamp - AI Essentials for Work	15‑week practical course to learn prompt writing and AI at work	Nucamp AI Essentials for Work syllabus (15-week bootcamp) • Register for Nucamp AI Essentials for Work
Grant support	Seed funding and workshops to scale AI‑cyber training across the Inland Empire	Rep. Pete Aguilar $300,000 award to CSUSB press release (May 13, 2025)

“As artificial intelligence technology develops, it is crucial that the cybersecurity workforce in the Inland Empire is equipped with the skills and knowledge necessary to ensure we can keep up.” - Rep. Pete Aguilar