Within retail, there is much debate around whether unmanned stores and AI can potentially replace frontline labour. But the experience of Alex Ng, co-founder and CEO of Intellify, has led him to a different conclusion: AI may be more useful in removing routine interactions from human employees so they have more time for the customers that matter most.

Ng’s own thinking changed when Intellify attempted to apply the model it had developed for unmanned convenience retail stores to Singapore’s Mint Museum of Toys. The assumption was that if customers could enter and navigate a store without staff, the same approach could be applied to an entire museum. It quickly became clear that the two environments had a fundamental difference.

“In a convenience store, all the products are self-explanatory, the customers know exactly what they want, and they do not need any kind of human touch or explanation,” Ng says. Museums, by contrast, are built around discovery, questions and engagement. The challenge was not how to remove the human, but how to make the human more useful.

Different customer interactions carry different values. Answering a routine question about opening hours, product availability or how to buy a ticket may require little human judgement. Helping a customer understand a complex product, recommending between alternatives or turning a purchase into a more meaningful experience is different.

At the museum, AI took on around 80 per cent of the mundane and repetitive questions, leaving the employee to focus on customers with more involved requests, learning journeys and guided tours. During peak periods, the employee could use the AI as support; outside those hours, the AI could continue handling basic enquiries.

In another example of successful implementation, Intellify helped a premium snack brand with a retail store at Singapore Airport to trade 24-seven. Traditionally, the store closed from midnight through 6am, but Intellify created an AI-driven solution that allowed the store to continue trading without an attendant during the six-hour break. “They managed to capture this invisible revenue that was previously not possible without the need to add any additional manpower,” says Ng……

In another location, an AI concierge assistant was installed in a Singapore hotel. Data from its exchanges with customers showed that non-English speakers were spending two to three times longer talking to the AI compared with English speakers. “This gave us insight that non-English-speaking visitors or guests actually need a lot of engagement. But that was a big gap without AI because it’s so tough, or almost impossible, to find any employee who could speak multiple languages.”

Intellify’s data shows that language barriers are a big pain point for retailers. “People just try to live with it because if someone cannot speak your language, there is nothing you can do about it. So, with this AI implementation, we saw how the gap is being bridged with so many non-English languages being exchanged, and that was very insightful.” 

Ng does not envisage a future in which stores are unmanned around the clock. Instead, he expects a hybrid model in which human employees concentrate on peak periods, high-value customers and complex interactions, while AI handles routine engagement and extends the hours in which customers can receive assistance.

Ng says this will make human employees a lot more valuable, because they will be engaging with VIP customers and closing important sales instead of just answering generic questions about a product. That allows the retailer to deliver a more premium experience and play a really different role in getting that bit closer to the customer.

The importance of trust and honesty

Ng agrees that there is a significant gap between AI that recommends and AI that executes without human interaction. That is where the need for customers to trust the retailer comes into play. 

“Trust comes from both sides. From the customer’s point of view, approaching AI or using AI is definitely not the first instinct because that is not natural behaviour. How AI is being interacted with is not a norm yet; we are still at a very early stage, so humans are still required. An AI ambassador usually acts as a backup or a secondary alternative, so when they see a human that is engaged, at least they have AI that they can approach and speak to. 

“We think that trust has been increasing because we see the utilisation rate across different industries, from retail to hospitality to tourism, and even trade shows. We see how the engagement in terms of the duration spent speaking, and we saw how the funnelling took place to understand about the products and services that were asked about increase over time,” Ng continues. 

“With AI being able to provide accurate information and answer directly what the customers are looking for, that would help a lot, unlike a chatbot, which would probably take you round and round before answering your question, or maybe it would would not even answer your question at all.”

For retailers and other organisations adopting this AI technology, honesty is important. They should look at the data and measure how the solution is performing, the insights being captured, and what things are not being done well. 

“These are the kind of refinements that need to take place along with the technology, together with the operations of the organisation. That honesty allows the organisation to progress with using such AI tools to make their operations even better.”

Such solutions are a learning process; you need to be committed to tweaking, changing, and activating what the AI knows, concludes Ng. 

• Listen to the podcast to hear Ng project the future of autonomous retail and AI-powered customer engagement during the next three years, what advantages early adopters of the technology will have over those who wait and why he believes the era of giant big-box stores is over.