Sandeep Shekhawat, Director of Engineering at Walmart Global Tech (USA), explores how a fusion of AI and IoT technologies enables retailers to design, test and deploy innovative new customer experiences.
Retailers looking to stay one step ahead of their competitors are turning increasingly to artificial intelligence (AI) and the Internet of Things (IoT), using these technologies to help them understand the millions of data points in stores and turn them into usable insights that help improve the speed and effectiveness of their business decisions. The retail sector is leading spending on AI across industries, reaching over $4.5 billion in 2021 and predicted to grow up to $31.98 billion by 2028.
The rapid development of new technologies has led to many changes in retail, a space where recent years have not only seen customers shift between physical and online shopping, but where the COVID-19 pandemic saw retailers necessarily adapt quickly towards click and collect services. They have also invested in touchless experiences and are leveraging IoT technology to enhance customer experiences.
Retail is, then, a sector in seemingly constant transformation.
The convergence of AI and IoT (AIoT) opens significant opportunities, especially for brick-and-mortar stores. Stores now use a huge number of sensors to collect data about customers. Increasingly, this data can be processed within the sensors themselves, reducing latency and cloud storage round-trip times. By deploying sensors, smarter controls and increased sources of data collection (from both business and customer), retailers can fully take advantage of these technological capabilities.
This fusion of AIoT in the retail industry creates many opportunities. These include:
- Refining operations to increase store efficiency
- Driving more productivity
- Creating a better, more personalised shopping experience for customers
The technology is also used in workforce management, planogram management (placing goods on shelves) and inventory fulfilment. Elsewhere we’re seeing computer-vision shelf scanning and cashierless checkouts. These uses of AIoT are driving significant investor interest back towards physical retail.
Why should we use AIoT in retail?
The most common applications of IoT devices in retail - connected over wireless networks or Bluetooth – fall into three interlinked, high-level categories:
- Customer-facing experiences
- Operational excellence
- Lifecycle management
According to Gartner’s predictions, 80% of IoT projects will have featured an AI component by 2022 and the combination of AIoT is expected to expedite the digital transformation of various industries globally. According to a report by the SPD Group on The Value of Artificial Intelligence for Retail in 2022, one AI vendor claimed a 32% reduction in operational costs after implementing AI-driven solutions.
Customer facing experiences
‘Customer experience’ focuses on the relationship between a business and its customers. Looking specifically at retail, today’s customers want to be in and out of the store quickly. Some examples of AIoT driven elements which drive a seamless, quick shopping process are:
- Personalised experiences through apps and online using data-driven machine learning (ML) based recommendations to suggest products, prices, and promotions to customers while they’re online or in a store
- Cashierless checkouts were a huge benefit during the pandemic because they reduced contact between people. Since then, retailers have rushed to install self-checkouts and ‘Scan ‘n’ Go’ experiences, such as Walmart+ Scan n Go, which can be a low-cost way to implement cashierless checkouts
- Vision-based product search tools which use AI/ML and trained models to search for products within an app. Once the product is found, the system allows the user to try the product using augmented reality (AR)
- Streamlined in-store navigation and heat map systems make it easier for customers to locate products, using IoT to show augmented reality (AR) based directions in the store. This allows speedier in-store navigation, fewer staff interactions and increased purchase rates. In essence, this provides customers with a digital shopping experience even in a physical store
- AI/ML based automatic call centres and customer support which reduce the load on in-store staff during holiday and peak shopping periods by enabling natural, personalised interactions with virtual agents
Operational excellence
Retailers work on thin margins and in a hyper-competitive labour market, where customer loyalty is limited and supply chains are often unpredictable. Some examples of how retailers can use AIoT to drive operational performance and generate real impact are:
- Alarms and notifications
Measuring temperature, humidity, and lighting in specific store areas can help reduce the wastage of frozen and fresh produce - Inventory management
AI can optimise inventory by using computer vision systems and object detection software to track inventory and send notifications for replenishment even when facing supply chain disruptions - Improving business operations
Door security sensors gather information and help monitor and control other equipment including emergency lighting and store systems such as security and air conditioning. Smart systems can also be used to track changes in energy consumption, system availability, and performance across the day and during emergencies. All these activities generate data which can be analysed to gain usable insights which help improve operations - Data analysis
IoT devices collect large amounts of data, but real-time processing can be challenging. AI and ML-trained models can be deployed to detect fraudulent transactions and make real-time decisions on customers’ historical buying habits as well as suggesting promotions and offers
What is a digital twin and how can they be used in retail?
A digital twin is a virtual replica of a physical process, product, or system, from manufacturing processes and supply chains to infrastructure. They allow organisations to simulate, analyse and test different scenarios and approaches.
For you
Be part of something bigger, join BCS, The Chartered Institute for IT.
In the retail industry, digital twins can be utilised to optimise different processes and improve decision-making. For example, a digital twin of a store's inventory system can test different stocking and replenishment strategies, allowing retailers to identify the most efficient approach. Similarly, a digital twin of a store's layout can be used to model different merchandising and display strategies, helping retailers to enhance the customer experience.
They can also improve pricing accuracy and reduce losses caused by incorrect pricing. Using real-time data from in-store sensors and other sources, a digital twin can track and update product pricing in real-time, helping retailers avoid costly errors and lost revenue.
The benefits of using digital twins in the retail industry are numerous; they allow retailers to test different scenarios and strategies without incurring the costs and risks associated with implementing changes in the physical world. They can also monitor and analyse customer behaviour, allowing retailers to understand their customers' needs and preferences. Additionally, digital twins can help retailers to optimise their operations, financial performance and decision-making by providing real-time data and insights.
Conclusion
AIoT provides insights into customer preferences and allows for cost-cutting through efficient operations, resulting in competitive pricing and a better in-store experience for customers. Retailers must embrace it to stay competitive: both large and small businesses can benefit from this technological revolution in retail.
