NeuraIQ
Home
AboutBlogContact
Book a Demo
NeuraIQ

Edge AI That Grows Your Business, Without Giving Up Your Data

Australian based technology company specialising in privacy-first Edge AI solutions.

🇦🇺 Sydney, NSW, Australia

Products

IntelliGuardNeuraMateNeuraHomeNeuraBotsNeuraSitesNeuraApps

Company

AboutBlogContactRequest a Demo

Contact

admin@neuraiq.com.au0434 673 004neuraiq.com.au
© 2026 NeuraIQ Pty Ltd. All rights reserved.
Privacy PolicyTerms of Service
← Back to BlogIndustry

AI Theft Detection for Australian Pharmacies: What Actually Works On-Premises

By Mina George  ·  July 2026  ·  14 min read

Saturday, fragrance bay. The pharmacist is counselling at the dispensary. A customer opens a tester, then a second bottle goes into a lined bag. Playback on Monday will show the event. By then the stock is gone and the only decision left is write-off versus insurance.

That is the gap AI theft detection is supposed to close: move from post-event evidence to in-aisle intervention. The market is crowded with claims. Australian pharmacies need a sharper filter - what the crime data actually shows, how the pipelines work, and which architecture choices create Privacy Act and latency risk.

> Quick answer: For Australian pharmacies, effective AI theft detection means real-time concealment analysis on existing CCTV, with alerts fast enough for floor staff to act, preferably with inference on-premises so customer video does not leave the building for routine processing.

At a glance

  • →PPFF / Griffith 2022: ANZ crime-related retail loss ~0.88% of revenue (~AUD 4.3 billion sector-wide); external theft alone ~AUD 2.29 billion.
  • →Pharmacy hot products: perfumes, facial creams, make-up; booster bags disproportionately reported in pharmacies.
  • →Median customer-theft apprehension value in the study: $415; most incidents involve two to five items.
  • →Gesture / behaviour models differ from facial recognition - demand the distinction in the contract.
  • →Edge inference: milliseconds to ~2s alerts; cloud upload paths: often 5-15s and APP 8 exposure if frames go overseas.
  • →Success metric is shrink prevented and staff-usable alerts, not demo-reel accuracy slides.

In this article

  • →What does the Australian crime data say?
  • →What products and methods hit pharmacies hardest?
  • →How does AI theft detection actually work?
  • →Gesture detection vs facial recognition
  • →Why latency and architecture decide outcomes
  • →Privacy Act, CCTV, and APP 8
  • →What traditional controls miss
  • →Vendor evaluation checklist
  • →Where IntelliGuard fits
  • →The bottom line

What does the Australian crime data say?

Skip global “retail theft is rising” headlines. Use regional numbers you can defend in a banner-group business case.

The 2022 Australia and New Zealand Retail Crime Study, produced by Griffith Criminology Institute researchers for the Profit Protection Future Forum (PPFF), surveyed companies operating more than 8,900 stores with ~AUD 136 billion turnover - roughly a third of the ANZ industry.

Key findings from that study:

  • →Average crime-related losses for FY2021-22: 0.88% of revenue, extrapolated to about AUD 4.3 billion across the ANZ retail sector (up ~28% in absolute terms since FY2017-18’s AUD 3.37 billion estimate).
  • →Customer theft was the largest crime-related loss category at 53%, ahead of employee theft (24%), customer fraud (14%), and vendor fraud (8%).
  • →Estimated external theft cost for the sector: AUD 2.29 billion.
  • →Typical external incident: two to five items; median value of a customer-theft apprehension: $415.

Those figures are not pharmacy-only. They are the sector baseline. Pharmacies sit inside that distribution with a distinctive product and method profile (next section). For store-level cosmetics economics we use elsewhere, see the real cost of retail shrinkage in Australia.

Recruiting context matters too: 94% of respondents in the same study said hiring LP staff with relevant experience was somewhat or extremely difficult. Independents rarely have a dedicated LP analyst. Tools that create actionable alerts for floor staff matter more than tools that create more footage for an empty LP seat.

What products and methods hit pharmacies hardest?

The PPFF study’s pharmacy hot product ranking for the period was:

Pharmacy hot product (PPFF 2022)Why it is targeted
Perfumes and fragrancesHigh resale value, small form factor, open-shelf display
Facial creamsPremium unit price, easy bag concealment
Make-up productsHigh volume SKUs, hard to spot missing units until stocktake

That matches what pharmacy operators describe operationally: small, high-value, open-shelf SKUs with strong secondary markets. Industry explainers aimed at pharmacy also commonly list infant formula, vitamins, razors, and some OTC packs as frequent targets - useful for zone planning even when they sit outside the PPFF top-three list.

On methods, the study found most external theft used low-sophistication approaches (concealment, walking out unpaid). Planned techniques were category-specific. Booster bags were reported most commonly in pharmacies, then apparel - a signal that pharmacies should treat bag-lining and foil/metal-lined bags as first-class detection scenarios, not edge cases.

Distraction techniques also rose across categories versus the 2019 survey. AI that only watches a single camera crop without multi-person context will miss the classic “one distracts, one conceals” pattern.

How does AI theft detection actually work?

Strip the marketing. A production pipeline looks like this:

  1. 1.Ingest RTSP / NVR streams from cameras already covering high-risk bays.
  2. 2.Detect and track people across frames (not identity - bounding boxes and track IDs).
  3. 3.Pose / hand trajectory estimation for pocket, bag, clothing, and shelf interactions.
  4. 4.Action / gesture classification for concealment patterns (and optionally product sweeping).
  5. 5.Temporal scoring so a single ambiguous frame does not page the whole floor.
  6. 6.Alert delivery with clip + location to phone, tablet, or floor display.
  7. 7.Operator feedback (true / false / ignored) to tune per store.

Vendors differ on where steps 2-5 run (edge appliance vs cloud GPU), how many cameras are licensed, and whether alerts include face crops for staff approach. Those choices dominate privacy and latency more than the brand name on the box.

For a deeper walkthrough of a multi-model on-premises cascade, see how AI detects shoplifting in real time. For why false positives kill adoption, see what is a false positive and why it matters.

Gesture detection vs facial recognition

Australian pharmacies are health-adjacent environments. Customers expect discretion. That is why several commercial systems (including widely marketed gesture products) emphasise behaviour and body movement, not facial biometrics, for primary theft detection.

That distinction is load-bearing:

  • →Gesture / pose models ask: did a concealment action happen?
  • →Facial recognition / re-identification asks: who is this person, and have we seen them before?

Both can appear in one product. Some stacks use face crops only as a staff visual for “who to approach”, stored briefly or as embeddings under staff-controlled policies. Others market chain-wide offender matching. Those are different legal and cultural products. Put the difference in the statement of work.

OAIC guidance is clear that security-camera footage can contain personal information when individuals are reasonably identifiable, and APP entities must notify, secure, and destroy or de-identify that information when it is no longer needed. Facial databases amplify that obligation. Gesture-first detection does not remove Privacy Act duties - it reduces the blast radius of what you collect for the primary purpose of loss prevention.

Why latency and architecture decide outcomes

A 2026 Australian industrial comparison of edge vs cloud AI puts the practical split plainly: edge inference lands in milliseconds for real-time work; cloud round-trips often land in hundreds of milliseconds to seconds, depending on model and network. Loss-prevention vendors who stream frames for remote inference commonly quote multi-second alert paths once upload, queue, and push are included. One pharmacy-facing vendor publicly cites roughly seven seconds detection-to-alert on their stack - still useful versus Monday stocktake, but a different intervention window than sub-two-second local inference.

Australian operating realities push edge further than US metro demos imply:

  • →NBN upload is the constraint, not download. Continuous HD multi-camera upload is expensive and fragile.
  • →Regional / suburban outages blank cloud-dependent detectors for the exact busy periods you care about.
  • →Streaming cost at 8-15 cameras, 24/7, is a P&L line, not a footnote.
DecisionCloud / hybrid uploadOn-premises edge inference
Where frames are analysedVendor data centre (often overseas)Appliance in your store
Typical alert pathUpload → queue → infer → push (seconds)Infer locally → push (sub-second to ~2s)
APP 8 exposureCross-border disclosure of identifiable video may applyNo routine overseas disclosure of footage for inference
NBN outage behaviourDetection often stops when upload failsCore detection continues offline
Bandwidth cost at 8-15 camerasContinuous HD upload is materialOnly alerts / metadata leave the LAN if anything

Hybrid designs exist (edge for inference, cloud for dashboards). The question to ask is whether raw pixels leave the site for routine inference, or only structured events (alert IDs, timestamps, optional short clips under your retention policy).

Side-by-side detail: edge AI vs cloud CCTV comparison and why on-premises AI beats the cloud for pharmacy security.

Privacy Act, CCTV, and APP 8

This is not legal advice. It is the diligence pharmacies should run with their adviser.

  • →Coverage: The Privacy Act 1988 and Australian Privacy Principles apply to many organisations with turnover above $3 million, and to certain categories (including many health service providers) regardless of turnover. Confirm your entity status.
  • →CCTV as personal information: OAIC’s public guidance treats security-camera images of identifiable people as personal information subject to the APPs (notice, security, retention).
  • →APP 8 (cross-border disclosure): Before disclosing personal information to an overseas recipient, an APP entity must take reasonable steps to ensure the recipient does not breach the APPs, and remains accountable for mishandling (Privacy Act s 16C), subject to limited exceptions such as informed consent. Uploading live CCTV to an overseas AI service is exactly the scenario APP 8 contemplates.
  • →APP 11 (security): You must take reasonable steps to protect personal information. “The vendor’s DPA says so” is weaker than “the pixels never left the LAN.”

On-premises inference does not make you exempt from the Privacy Act. It removes a large class of overseas-disclosure and bandwidth failure modes that cloud-first video AI introduces by default.

What traditional controls miss

EAS gates, locked cabinets, and CCTV playback still have roles. They fail the pharmacy pattern in predictable ways:

  • →EAS catches some packaged goods at exit - not open testers moved into a personal bag mid-aisle.
  • →Locked cabinets protect SKUs and destroy browse conversion on cosmetics.
  • →Guards are expensive and scarce; the PPFF study already flags LP hiring difficulty.
  • →Playback documents loss after the booster has rotated to the next suburb.

Organised and semi-organised crews exploit time: seconds in the bay, minutes to the carpark, days before stocktake. Real-time detection only helps if the alert arrives inside that first window and staff trust it enough to walk over.

Vendor evaluation checklist

Use this in demos. Score vendors in writing.

  1. 1.Inference location - appliance in store, on NVR, or cloud GPU? Ask for a network diagram.
  2. 2.What leaves the site - raw frames, clips, embeddings, or metadata only?
  3. 3.Data residency - if anything syncs, is it Australian-hosted? APP 8 paperwork ready?
  4. 4.Gesture vs face - which models fire the primary alert? Retention for face crops?
  5. 5.Camera compatibility - RTSP brands, max streams, resolution / FPS requirements at shelf height.
  6. 6.Latency SLOs - published p50 / p95 alert times under load, not best-case demo.
  7. 7.False-positive handling - how do staff mark noise, and how fast does per-store tuning land?
  8. 8.Booster-bag / multi-person scenarios - show pharmacy footage, not supermarket self-checkout only.
  9. 9.Offline behaviour - what works when WAN drops?
  10. 10.Commercial model - per camera, per store, hardware included? AUD pricing excl. GST.
  11. 11.Support - Australian hours, who calibrates zones after fixture changes?
  12. 12.Pilot design - one highest-shrink store, 30-60 day baseline, agreed kill criteria.

If a vendor cannot answer (1), (2), and (6) without marketing adjectives, you are buying a slide deck.

Where IntelliGuard fits

IntelliGuard is NeuraIQ’s managed, on-premises AI theft detection service for Australian retail and pharmacy. Inference runs on a NeuraIQ edge appliance in the store. Staff get fast face-photo alerts and optional floor display cues so the whole team - not only the person at the NVR - can respond. Video is not uploaded to cloud servers for routine inference.

Commercial shape: subscription pricing from $299-$369 AUD per store per month (excl. GST) depending on store count, appliance included, 30-day money-back guarantee. Pilot the highest-shrink site, measure incidents interrupted and dollars retained, then expand on evidence.

Book a walkthrough: request an IntelliGuard demo.

The bottom line

AI theft detection is not “smarter CCTV.” It is a real-time intervention system sitting on top of cameras you already own. Australian evidence says pharmacies lose margin on fragrance, skincare, and make-up to methods that include booster bags and distraction. Systems win when they (a) detect concealment gestures in seconds, (b) keep identifiable video processing local enough to satisfy operational and APP 8 reality, and (c) produce alerts staff will act on.

Demand architecture answers before brand stories. Then measure shrink on the floor - not accuracy on a demo reel.

Want to see it in action? Book a 15-minute demo and we'll show you IntelliGuard detecting concealment on a live camera feed.

Book a Demo →