The retail business has many stakeholders.
We have an entire spectrum of retail space providers. We have large retail chains like Walmart and Dmart in India, and we also have your neighbourhood mom-and-pop stores or the paan-beedi khomcha
Another stakeholder is the entire supply chain ecosystem. They range from traditional localized distributor-based models to semi-centralized hub-spoke warehouses to fully centralize the D2C enablers.
The third stakeholders are the Consumer Packaged Goods (CPG) companies. They include the product planners and designers, the manufacturers, product category managers and merchandisers.
Lastly, but most importantly we have consumers. The consumers may have different paying capacities. They can be adventurous when it comes to choosing products or be traditional and stick with their preferred brands.
The above mentioned stakeholders are the decision makers of how and to what extent artificial intelligence will be incorporated into their retail shopping experience. Which means that the future of AI in retail, rests on the benefits AI provides to these stakeholders.
The Large Retail Chains
Out of stock (OOS) causes massive revenue leakages for large retail chains. A typical supermarket has 15,000+ SKUs while a convenience store has 5,000+. Ensuring the constant availability of all these products is a daunting challenge. Retailers typically experience OOS levels of up to 30 percent of SKUs, depending on the category. This leads to a ~4 percent loss in sales for a typical retailer and globally adds to $1tn+ in lost sales as per a recent study. This is where the AI steps in, and in so many ways :-
• AI-based Shelf Monitoring – AI solutions like ShelfWatch can be used to ensure that shelves are replenished on time and the most basic OOS losses don’t take place. Having a product in-store inventory and a customer not being able to buy it due to delays or mistakes in restocking is the most basic problem retailers should start addressing with AI.
• Forecasting – New, more accurate AI-based forecasting methods can predict demand for various products depending on weather, holidays, government policies, locality, and other variables that influence sales. Better forecasting avoids not just OOS due to stock shortage but also improves the bottom line by avoiding overstocking.
• Personalization of Shopping Experience – While an online store like Amazon can quickly use data to personalize offers and prices for customers logging in, eventually AI will enable smart CRMs which learn from customer’s past behaviour and current data to try and suggest cross-selling and upselling products.
• Ensuring Compliance – There are many compliances that the retailers have to follow – some products need to be kept at a required temperature range, hygiene compliances, COVID security protocols, and more. AI can help retailers make sure these policies are not being breached at ground level. Apart from that, for retail execution purposes, AI can help in shelf monitoring to ensure KPIs like planogram compliance, eye-level placement of high salience products, right promos and price displays etc., are all met.
The Small Retailers
When it comes to small retailers, AI can help in many ways-
• Inventory Management – Small Retailers (especially from developing countries) unlike large retailers, cannot invest in costly inventory receipt and inventory management systems and associated staff. As a result, most inventory management is feedback and intuition-based. As in, when an item is low in inventory or an OOS is seen, the shopkeeper tries to make the best guess of how fast the item sold and replenishes stock accordingly. An AI solution can digitize the whole process which presents the true picture.
• Lack of a Receipt System – It can lead to other problems like unaccounted thefts from warehouses. Large retailers have in-house analyst teams who can use a combination of data/experience and suggest the right products to stock. AI will become the great equalizer here as it automates the sales-inventory analysis workflow and creates inventory suggestions automatically.
• Payment Infrastructure – POS systems and credit card machines are a big cost for many mom-and-pop stores as well. An AI + cloud enabled POS system could make such technology ubiquitous.
• Self-Checkout – While RFID-based self-checkout systems are being tried on by supermarket retailers so that shoppers spend time shopping and not in a billing queue, for relatively small retailers, such expenses are not possible. AI will bridge the gap as a combination of cloud and image recognition can be used for self-checkouts.
The CPG Companies
The CPG manufacturers are currently using AI to perfect their retail execution.
• Perfect Retail Execution – CPG creates guidelines to measure key performance indicators like planogram compliance, price display compliance, on-shelf availability measurement. Through image recognition solutions, field reps capture images of retail store shelves. These are fed to the AI which calculates the KPIs in real-time and instant action is taken to address the gaps.
The Supply Chain Players
In case of the supply chain players, AI can aid in-
• Reducing Misdeliveries - Misdeliveries are one major reason for leakage for supply chain providers. Many misdeliveries are due to the wrong allotment of packages - when a delivery person is assigned a package that is not in his service area. Here, the AI in combination with GIS systems can parse the address lines and match the packages to a point on the map. It can then allocate the package according to the service area of the delivery person.
• Automatic Warehouses - Image Recognition and Robotics can automate and speed up the working of warehouses. They can perform automatic sifting according to requirements of individual stores and creating delivery parcels ready to be picked up.
It's keeping in mind the customers that shopping experiences are designed. AI works proactively here as well-
• Better Products: AI can be used to analyse product usage to create more ergonomic and useful versions of products. A short explainer video can be found here.
• Better Prices: AI can process product purchase data and optimize production/retail inventories to make sure customers get the best prices. The pricing, design processes are still chiefly intuition-based and there are experts that design product/marketing campaigns. The more purchase/customer usage/customer behaviour data is available and processed by AI, the more effective these campaigns are going to become.
The ideas I have mentioned here are something that are already being implemented or will get wide adoption in the near term. When we speak of 2025 - most likely we will be shopping in self-checkout stores (similar to AmazonGo). Thus, AI is an all-pervading technology, and the work it achieves for the retail industry is just one of its paradigms. AI is multi-dimensional and can be integrated with other technologies to create robust systems with high ease of use. Its exactness and insightful observations are what makes it trustworthy to people. With the wonders it has performed in e-commerce, the stakeholders in the retail sector have now come alive to implementing this technology in retail as well.