Share This Page

Author: Arka Bagchi |03/13/18

Artificial Intelligence - The Answer to Retail Shrinkage

One of the top few concerns of physical retail in today’s world is Shrinkage. Shrinkage in retail is nothing but simple loss of inventory due to employee theft, shoplifting, administrative error, vendor fraud and damage in transit or in store.

Average retail shrinkage accounts for 2% of the sales. As per National Retail Security Survey, the value was more than $49 billion in 2016. The shrinkage rate in supermarkets is usually higher at 2.5% of the sales which is more than double of the regular 1.1 percent rate for all other retailers

Artificial Intelligence in Retail, Reducing Retail Shrinkage

In this blog, we have identified top five types of retail shrinkages and we will be discussing how artificial intelligence can help retailers that are encountering shrinkage. Towards the end, we will see how an AI company is mitigating retail shrinkage in real life.

Reducing Retail Shrinkage with AI

1. Non-Scanning Loss

Business Problem:

  • Identifying the products that go without scanning at Point of Sale (POS) check out
  • POS operators making mistakes while entering the count manually or by scanning

How can AI improve the process?

  • By applying advanced video and data analytics to existing POS video and data streams
  • Close-circuit TV and electronic POS data streams are fed to the Data Collection Unit which has dedicated on-board technology
  • The electronic POS can always be connected to the strategically placed surveillance system to track the entire POS activity and identify potential irregular operations

2. Basket Based Loss

Business Problem:

  • Customer fails to take out product from trolley to present at POS
  • The three types of basket-based losses are bottom of basket, middle of basket and top of basket

How can AI improve the process?

  • Computer vision system which is capable of understanding the full set of fraudulent behaviors
  • Web video streaming technology with video-to-transaction log synchronization
  • Video intelligence and image processing & recognition algorithms to immediately detect items left anywhere in the cart (Bottom, Middle, Top)

3. Sweet-hearting Loss

Business Problem:

  • Unauthorized giving away of merchandise without charge, to a “sweetheart” customer (e.g., friend, family, fellow employee)
  • Sweet-hearting costs the retail industry an estimated $14 billion annually

How can AI improve the process?

  • Identify variety of sweethearting behaviors:

a) Covering Bar code

b) Stacking items one on top of the other

c) Skipping the scanner entirely and directly bagging the merchandise

  • By applying advanced computer vision algorithms to the existing camera feeds, the system can see and understand what is going on at the checkout, and track every item
  • Associate all items within a transaction to the transactional data feed from the POS system to flag anomalie

4. Self-Checkout Loss

Business Problem:

Below are the few types of self-checkout losses faced by the retailer

  • Direct to bag
  • Price-Look-Up (PLU) abuse by tempering codes
  • Un-scanned items left in cart/basket
  • Bypassing the belt

How can AI improve the process?

  • An intelligent web interface that allows store associates to identify suspicious behaviors in real time, and take actions
  • Computer vision system that automatically identifies the mismatch between transaction receipt and the item being scanned in the video

5. Poor Demand Planning Loss

Business Problem:

  • Fresh products accounting for up to 40 percent of grocers’ revenue and one-third of the cost of goods sold
  • Since the products are perishable, demand is variable and lean times are uncertain; accurate replenishment planning is the key

How can AI improve the process?

  • Automating produce replenishment decisions offers retailers with more options for optimizing their fresh produce stock
  • Decision based on historical data is not effective. Decisions based on algorithms that allow computers to “learn” from data even without rules-based programming is going to be the game changer
  • Automatically including events like promotions into its calculations
  • Understanding the exact impact of over and under ordering at each individual store, and calculating which store can optimize the sales of available stock in hand

What is Standard Cognition and how is it changing the game?

Standard Cognition is an AI platform that enables buyers to shop in a cashless retail store by using computer vision and deep learning.

This AI enabled solution automatically identifies the product details picked up by a customer from the shelf and charges them while checking out and updates the inventory as well. It also understands if the customer puts the item back in the shelf.

The shopper app from standard cognition helps the buyer to just walk into a store, pick up the product and walk out of the store while payment is done from the digital wallet without any human touchpoint. Thus, providing a seamless retail experience to the customers.

The seller/store app helps the store staff to identify customer location and what they are buying. It matches the shopper with their corresponding basket and processes payment automatically while checking out. It also recognizes shoplifting and notifies the staff to redirect the customer to a kiosk or self-checkout counter.

As the boundary between physical and online retail keeps on reducing, it is high time to integrate artificial intelligence based computer vision solutions to encounter challenges like shrinkage. This will help the retailer to prevent loss, maintain the operating margin and focus on superior customer experience at the same time.

Read about Mindtree’s Intelligent Video Surveillance platform here. It is an open hardware agnostic platform, easy to host and provides security, safety and other compliance services based on video analytics.

Do you also believe that AI is the future of retail shopping? Let is know your views at


Let's Talk About Your Needs

Thank you for your submission. We'll be in touch.