Retail is a fast-moving, trend-powered industry that relies heavily on making smart predictions to drive the business.
Artificial Intelligence (AI) and Machine Learning (ML) can improve the accuracy of these types of predictions by incorporating a wider range of signals and finding patterns that often humans can’t. But successfully adopting AI and ML can be quite challenging.
In this final session for NRF’s Retail Big Show, Sachin Padwal – Director of Product Management and AI Solutions for Google discussed the Top 5 ways retailers around the globe are embracing AI to improve experiences for their employees and their shoppers.
Padwal began by discussing the shifting expectations of consumers. Customers now want real time recommendations, seamless omni-channel experiences, customised offers, order-ahead capability, and supply chain transparency. They want products in store and available when they want them.
While customers won’t ‘see’ AI, it will be present.
A good example, Padwal said, was ‘Magic Eraser’ on new Google Pixel phones. AI is used to not just remove the image of someone or something in a photo, but also re-construct the background behind them. While AI does this, all you see if the edited photo.
The First Way: Discovery
AI will not just help customers find products but find products they didn’t know they needed but are either good for them or provide benefits to them.
In the past, a customer might go online to find a product and they would be presented a list of alternatives, listed by companies who have paid to be listed first, then the most popular.
What AI will do is provide that customer personalised recommendations of only brands they normally buy, of products that are in-stock and located locally.
Personalised recommendations will be AI driven. So, rather that receiving an offer of a product they normally use, AI will predict what the customer is using the product for and offer ‘solutions’, based on individual preferences. AI will detect patterns in purchases, say arborio rice, garlic and chicken stock and make the prediction the customer is making risotto and then recommend an Italian wine, based on the customers individual tastes.
The Second Way: Conversation
AI is already enabling us to call someone while driving our car or change channels on our television. We ask SIRI and Alexa basic questions about the weather, or to turn our lights on or off. However, the language is disjointed and often unclear. AI will enable natural conversations, leading to faster service and meeting customers’ ever changing expectations.
Imagine standing in the fresh food department of a supermarket asking your ‘app’ to talk you through some alternative meal ideas? Or asking about an unknown ingredient in a can of soup.
Chatbot experiences will improve. Customers will no longer have to “say in a few words” – instead, the conversation will be natural, and solutions arrived at instantaneously.
Transactions will be easier. No more manually transacting an online purchase. With AI, you will simply say, “I’ll buy one of those, and please charge it and deliver it to my office tomorrow at 5pm”.
The Third Way: Inventory
Demand is local. Demand is not uniformed across countries, states, regions, or even suburbs.
For example, a sudden increase in sales of a few items may be driven by a local competitor having to close temporarily. AI will enable inventory systems to detect even the very slightest change in demand and identify patterns in behaviour that humans are unable to.
AI will assist in implementing real-time tailored assortment planning. Providing direction to store teams when to increase footage, location, even forecasting when to ‘order a little extra today’, because the AI system predicts a lift in demand tomorrow. Automated demand forecasting tools with real time insights
AI will help understand if the right inventory, is on the right shelf, in the right location, in the right quantity and at the right price, across all stores. AI will help share that data across all stores, so stores can optimise their ranges.
Add in a robot-driven cameras that can track inventory. The inventory robot will wander the aisles constantly, over the entire day, identifying when lines sell out and automatically re-ordering. The robot will also detect missing or incorrect price tickets, or products that have been filled into the wrong location.
The Fourth Way: Fulfilment
It’s always frustrating when an item is unavailable, but it is also more frustrating when the retailer is unable to give you a time when it will be back in stock.
Tied to inventory systems, AI will provide faster, more accurate, real-time information on fulfilment.
Delivery addresses will be validated, and routes optimised. Delivery drivers will be ‘routed’ and ‘re-routed’ through advance navigation systems. AI will scrape multiple sources of information, including traffic reports, maps, congestion and accident data to precisely provide delivery times to stores and customers. Data will enable the measurement of fleet performance.
The Fifth Way: Trust
Trust is vital in the rollout of AI systems. Anything that is built on AI, must be built on a basis of trust.
AI required significant volumes of data to work. Customer purchase and transactional data. Behavioural, voice, biometric data. Naturally, consumer education will be vital to allay fears related to personal information and privacy.
Retail leaders will need to carefully consider the important questions of fairness, ethics, transparency and governance, before investing in and operationalising AI systems.
