Generative AI in Retail Series
Verneek
May 19, 2023
Implementing Generative AI in retail can improve customer experience, assist in-store workers, increase shopper loyalty and drive incremental sales. Following best practices ensures the AI ensures broad adoption across shoppers and channels, Each stream within the omnichannel comes with its own friction, and understanding that friction is the first step in creating a Shopping AI. Although price is the lever most retailers pull to drive immediate sales, customers' constraints go way beyond cost, and the number and variety of issues shoppers consider are growing. Some of the information shoppers seek is fundamental, and some are more complex, but the success of Shopping AIs will depend on accuracy and transparency. Likewise, data collected from the shoppers must be handled with care for a Shopping AI to become a trusted advisor.
Understand the Friction in the Process within each channel
In the previous blog, we discussed the friction in shopping and how Generative AI can help customers navigate their preset constraints ( time, resources, personal conditions) to quickly discover product services and solutions to meet these often conflicting criteria.
Today, the digitalization of the shopping process is transacted in many different places, and the channels continue to expand. The smartphone enables Text to Shop, Digital Assistants, and other on-the-go shopping experiences. Much of the research shows people still enjoy the brick-and-mortar experience of viewing selections, making comparisons, and sparking ideas. For its ease, the convenience of online shopping is compelling for many customers strapped for time that need to get some basics quickly. Each of these channels has its advantages but comes with some inherent friction.
Going deep into the friction facing consumers by channel provides a more precise road map for Shopping AIs to assist the customer in making more intelligent and faster decisions. Knowing the specific friction points for your target shoppers is vital to designing an experience for that channel. In our over-abundant world, product and assortment proliferation alone can confuse shoppers and, depending on the category, could be the most significant constraint, often freezing them from making a purchase, fearing they will make the wrong choice.
Moving beyond Price
Almost 100 percent of the communication to potential shoppers across retail is information on price, promotion, or couponing. The weekly flier and the Sunday paper advertisements emphasize the hot price. These promotions do work for some. They induce people to act for fear of missing out on a great deal, and the sense of urgency drives action.
When you shift the perspective to the actual shopper's issues, price is only one constraint and often not the primary point. For a mall shopper, the biggest problem may be the bus schedule or being able to park safely. Apple understands that price is often not the main barrier to a sale. You will never see someone spinning a sign outside an Apple store promoting savings or large price savings signs in their window. They are not ignoring the price-conscious consumer. Apple has a website for refurbished products, but they don't highlight or advertise its existence. They know shoppers with price as their primary constraint will find their way to that site independently.
The questions and friction shoppers experience goes way beyond the price. Retailers should be prepared to include the cost in the Shopping AI solution but know that pricing is one of many attributes customers navigate.
“What is a healthy kids snack under $5?”
“What is a top-rated paella for 4 under $30?”
Getting foundational data correct
In the early 20th century, a farmer could call the general store and speak to the owner to verify the product they need is in stock before making the long trip into town. With all the capex expenditures on supply chain, inventory management, and optimal display sets, today's shopper struggles to get an accurate view or answer on the in-stock situation or aisle location for a specific product in a store. Suppose a shopper makes a trip to a store based on incorrect inventory information. In that case, it costs the consumer time, money, and energy, impacting loyalty and possibly the shopper's lifetime value. It may seem basic, but consistently providing precise inventory information and a transparent aisle location to the omnichannel shopper and the in-store worker are fundamental building blocks in successful AI retail implementations.
If your shopping AI can not provide precise in-stock and aisle locations, shoppers will never explore the higher-bound abilities that LLMs can deliver.
“What aisle is the mayonnaise on?”
“Do you have any non-alcoholic beer in stock?
Robust and Accurate Product Attributes
One of the strongest features of Shopping AIs is their ability to ingest large product and product attribute catalogs. A robust and highly accurate attribute set enables shoppers to navigate their constraints to find a product that meets their needs quickly. Voice-enabled Shopping AI can promptly sift through multiple restrictions or feature requests.
"I am looking for a gluten-free, sugar-free, vegan salad dressing under $5.00."
"Do you carry a gold watch with a light blue face in the 500-700 dollar range made in the US?”
As mentioned in the previous blog, these constraints are expanding as shoppers consider other concerns about where they spend their hard-earned dollars. A growing problem with many consumers beyond the price, ingredients, and packaging is the other ESG issues like carbon footprint, free trade, and sustainability. Like health or wellness scores, these newer concerns that shoppers consider when purchasing may not be called out on the package or in the product's online description.
HowGood, a company focusing on product sustainability, has over 100 metrics on 2 Million products. AI enables shoppers to ask questions like" What is the highest rated, lowest priced, most sustainable coffee available in this retailer?"
Ratings and Recommendations need to be impartial.
As the use of Shopping AI grows, the expectation of guidance and recommendations will become commonplace. Some of these will be based on the features of the product.
“What cough medicine will not make me sleepy?”
Other questions will be based on product ratings and other constraints.
"What is a top-rated wine to pair salmon under 30 dollars?"
The ratings and recommendations should be accurate and impartial for retailers to build loyalty and for shoppers to view the Shopping AI as a trusted advisor. Certifications and ratings that are paid for rather than earned undermines the credibility of AI recommendations and will limit the utility of AI. Shoppers may shop at retailers that provide more accurate and transparent data in their recommendations.
The Double Edge Sword of Personalization
As shoppers begin to use Shopping AIs in retail, there is an ability to learn and store the preferences and preset constraints that a shopper is navigating. Knowing a shopper needs to avoid a particular ingredient or material can help with future guidance from the AI. As AI begins to ingest these personal constraints, shoppers' friction is reduced. Shopping AIs will anticipate a shopping trip and even make recommendations that the shopper had not considered. AI can provide tailwinds to the customer journey as the AI evolves into a trusted advisor and sidekick to the process. As the consumer shares more data about their constraints and priorities, the more the Shopping AI and retailer can address their needs.
“Do you carry gum-free ice cream?”
“I am looking for a gluten-free, sugar-free vegan salad dressing?”
When many people in the industry speak about "personalization," they are often referring to more targeted advertising to an individual. Knowing a shopper or household has an allergy or preference for animal-safe cosmetics is valuable information for marketers of products that cater to those needs. The more information a shopper shares, the more this data can be used to target them for specific ads or promotions. Some of this activity may be welcomed, but much will not be, especially if it is evident that the offer was sourced from using the Shopping AI. Retailers utilizing Shopping AI must balance helping the shopper and building longer-term loyalty vs. using this personal data to target them for short-term sales.
The retailer's primary focus in offering a Shopping AI should be to reduce the channel friction for the customer, starting with some fundamental pain points and then expanding to higher-bound issues. The ROI in developing an authentic and transparent Shopping AI will come from a more loyal customer base and increased sales, particularly in complicated categories that the AI can make more accessible for the customer to shop.
