400 million users ask ChatGPT for assistance at least once a week. The app quickly became an advisor for the most pressing daily issues.
Retailers are also turning to AI assistants to address the shortage of store workers. AI assistants have already become commonplace in ecommerce of self-service portals. This year they are making big inroads into brick and mortar retail.
ChatGPT told me that…
True story – a customer in an electronics store in one of the larger cities in the Baltic states is purchasing headphones. To aid his decision making, he asks a human store assistant for help. Yet he checks every assistant suggestion with the ChatGPT app on his phone.
“ChatGPT recommended this model… Well, ChatGPT says that these headphones work better with Apple devices”, states the customer.
“Such sights are becoming commonplace in retail. ChatGPT offers a convenient way to ask questions when present at the store – without the need to google for reviews, specifications or customer feedback”, says CEO of retail technology solutions provider Neto Baltic Rokas Budvilaitis.
According to R. Budvilaitis, even if AI might hallucinate, the store assistant might also not know full specifications or edge aspects about the product.
“In this case, for the customer two opinions might be better than one”, adds R. Budvilaitis.
Virtual assistants for physical retail
Visitors to the retail technology trade show EuroCIS in Dusseldorf could test specialized virtual store assistants by themselves. One such assistant was working at a football boot section in a footwear store.
The assistant would help customers to find the perfect football boots for their use case – whether the shopper would play on real or on artificial turf; whether the purchase is for an adult or for a child; the assistant would discuss the budget and size availability at the store. At every step a human consultant could take over the process.
Although AI still follows a predetermined sales script, the customer can ask follow up questions (e.g. which boots are the most durable?) or ask the AI to justify his opinion (e.g. why are you suggesting this model in particular?)
“AI assistants can assist customers when buying more technically complex products, where simply outlying product specifications or providing simple product information at the store does not cut it”, explains R. Budvilaitis.
According to R. Budvilaitis, natural speech recognition and ability to communicate in natural language drives AI assistant adoption.
“While AI assistants can communicate in English or German, they cannot yet fluently speak in Lithuanian, Latvian and Estonian languages. Answers in these languages can be provided in text format on a screen. Yet in a few years smaller languages should become available”.
Does your store need a virtual AI assistant?
AI assistants can address shortage of store employees, reduce operational costs and help to boost sales. But – before adoption – retailers should define what customer journey issue they want to address.
“For example, at a typical footwear store, the most common customer question is whether the store has their shoe size in stock. Yet stock availability can be displayed on electronic shelf labels. Solum ESLs also have an in-built button to call the shop assistant. This is just one example where customer friction can be reduced with simpler and cheaper solutions”, says R. Budvilaitis.
An interactive product shelf could be another alternative. Neto Baltic developed and installed such solution at a specialized beverage store.
“When a customer takes a bottle from the shelf, product information – such as country of origin, taste notes etc. – are displayed on the screen. Computer vision is employed to identify which bottle was picked up. It augments the natural customer behavior; the customer does not need to “start a conversation with a screen” and hope for meaningful AI suggestions. And the most effective retail solutions are those that customers are actually using”, adds R. Budvilaitis.
AI assistants are the new SCOs
According to the research by Bain & Company consultancy, 41% would not mind using AI assistants in their purchasing process. Yet 57% of the respondents noted that they encountered AI mistakes and bad advice from AI assistants that negatively impacted their customer experience.
In the opinion of R. Budvilaitis, AI assistant journey into retail is somewhat familiar to self-checkout adoption.
“Self-checkouts made significant technological improvements over the years to become the preferred checkout method. And they are still improving to make them faster and easier to use. For instance, AI powered age verification is now installed at Partner Tech self-checkouts, reducing friction while purchasing age restricted items. Picklist Assistant automatically identifies produce and unpackaged goods without the need to look for them in the picklist menu, further increasing checkout speed”, says R. Budvilaitis.
According to R. Budvilaitis, after learning Lithuanian, Latvian and Estonian, as well as improving the quality of AI given answers, virtual assistants could become a viable alternative to traditional customer journeys.
