TDWI Articles

Executive Q&A: Applications of AI and Computer Vision in Retail

Computer vision takes sensor analytics a step further using AI. Rohan Sanil, CEO and co-founder of Deep North, an intelligent video analytics company, explains.

Upside: What is the difference between sensor analytics and computer vision in the retail environment?

Rohan Sanil: Sensor analytics is based on the use of physical sensor devices in-store. These work by converting stimuli such as movement or sound into electrical signals, which are converted to code and then processed by computers.

For Further Reading:

How Ethical AI Is Redefining Data Strategy

Are You Ready for Algorithmic Retail?

What Will Determine the Winners and Stragglers in the 2022 Digital Commerce Economy 

Computer vision is a field of artificial intelligence (AI) that focuses on replicating the capacities of human vision. As such, computer vision takes sensor analytics a step further because anything one’s eye can see can be analyzed. It trains computers to interpret and understand the visual world the same way humans do. For example, sensors may be able to pick up on whether a person has walked into a store and into a particular aisle. However, not only can computer vision can pick that up, it can also identify other events, such as whether there’s a dangerous spill in the aisle that could cause harm to a shopper.

How can AI and computer vision empower retailers to gain data-rich insights to better use their in-store environments so they can best meet shoppers’ needs and run their stores more efficiently and effectively?

It’s well known that traditional retailers have struggled to drive foot traffic and customer engagement in brick-and-mortar stores over the past decade. The Internet and the rise of retailers such as Amazon -- coupled obviously with the onset of the pandemic -- have changed the retail landscape in unprecedented ways.

With consumers now venturing back into stores, how can retailers effectively compete with their online counterparts -- and other brick-and-mortar businesses? A clear imperative is to deliver personalized service, convenience, and other engagement factors to drive purchases and loyalty.

What physical stores critically lack, unlike their online competitors, is clear visibility into consumers’ browsing and shopping activities. That could mean something as simple as how long a customer waits in line before making the purchase all the way to the in-store path-to-purchase for a shopper.

Brick-and-mortar retailers can use AI and computer vision in combination with their existing store camera infrastructure to understand who their customers are and how they behave while ensuring customer privacy. With these technologies, retailers can gain real-time insights for decision-making so they can positively improve key metrics such as in-store (and back-of-house) operations, labor planning and allocation, and -- critically -- overall consumer experiences. Retailers can do this through more effective product merchandising and marketing, staff optimization, and much more -- driving in-store conversions, satisfied customers, and significant cost savings.

What are the different areas in-store that these technologies can impact?

I’d offer five ways AI and computer vision can positively impact the in-store environment:

First, footfall analytics . With these technologies, retailers can determine metrics such as the number of people walking into a store, the number of people walking out of a store, and the total number of people in a store at a moment in time.

Second, customer demographics and repeat visitors . AI and computer vision can help retailers determine characteristics such as customer age range, gender, and the people who made more than one trip to the store on a single day.

Next would be the customer journey . With this functionality, retailers can understand heat maps or the number of entrances into a zone. They can also determine the length of time spent in-store as well as length of time spent in a specific store zone.

Fourth on my list is queue management and fraud prevention at checkout . AI and computer vision provide clarity on the number of people waiting in line for checkout and the average wait time spent in line before reaching the checkout. It can also reduce fraud at checkout and self-checkout.

Finally, in-store analytics . The technologies help retailers understand shelf engagement, including the number of touch gestures made towards shelved items. It also delivers POS transaction time and conversion details. In addition, they can provide retailers with insight into the dominant customer path, including zone-to-zone traffic patterns from shopper entry to exit.

What role can computer vision and AI play in retail loss prevention?

Retail shop margins are already under significant stress due to online competition. Furthermore, a retailer's bottom line might be harmed by a scarcity of inventory compared to the available records in the system, known as "shrinkage." Shrinkage is the loss of inventory stock as a result of shoplifting, employee theft, and other factors.

Shrinkage numbers leave merchants vulnerable because a higher amount of shrinkage means smaller profits. The National Retail Security Survey 2021 found shrinkage at an all-time high in 2021, accounting for 1.6 percent of a retailer’s bottom line.

Adding to that, in the post-COVID era, the desire to digitize the store and give more frictionless and contactless self-service, such as self-checkout alternatives, benefits both customers and retailers. However, with new advances comes the possibility of new channels for fraud.

To address these increasing challenges, retailers can use computer vision and AI to combat retail shrinkage with better loss prevention solutions at the front of the store. 

Fraud at checkout, including at staffed registers and self-checkouts, necessitates the integration of data from item-level tracking with computer vision and POS. Comparing item-level counts to POS-generated counts can help associates discover fraud and take action in real time. AI and computer vision can help retailers enhance checkout processes by making them smarter, which reduces theft and improves inventory control.  

In what areas beyond retail can computer vision and AI have an impact? How?

In addition to traditional retail, computer vision and AI can positively impact industries such as quick-service restaurants, shopping centers, transportation, commercial real estate, and manufacturing and warehouses. Across all these industries, the technologies can enable businesses to address a variety of challenges, including:

  • Limited visibility into in-store and back-of-house operations
  • Lack of insight into customer behavior and journey
  • Inadequate data for labor planning and allocation
  • Lack of real-time insights for decision-making

How is Deep North unique in this field?

Deep North is a pioneer in computer vision and AI, focused on providing insights for the physical world that previously could only be captured online. Key to our success are the patented and pending algorithms, proprietary pipeline, rapid deployment framework, and the actionable predictive analytics in the cloud about the customer journey, all built for enterprise scale.

Deep North helps brick-and-mortar enterprises optimize the performance of their physical locations by digitizing and analyzing behavioral metrics in the real world and providing the tools to act upon actionable, real-time insights while remaining compliant with CCPA, GDPR, and PII regulations.

[Editor’s note: Rohan Sanil is co-founder and CEO of Deep North. He is a business and technology executive with 20 years of strong product and business experience. You can reach Rohan via LinkedIn.

TDWI Membership

Accelerate Your Projects,
and Your Career

TDWI Members have access to exclusive research reports, publications, communities and training.

Individual, Student, and Team memberships available.