AI shop assistants: get ready for a world where you can't tell humans and chatbots apart

I regularly fly with KLM from Minneapolis to New Delhi, and always stop over in Amsterdam. I am frequently in Minneapolis for research and this is my route to go home to take a break from work. I have done the journey so many times that I know almost all the shops at Schiphol inside out. However, one time in summer 2019, the predictability was broken when I missed my connecting flight to New Delhi.

I was tired, hungry, sleepy, and the customer-service counter was closed. I had the choice to make the long walk to customer services at the next gate or use my iPhone, so I tried my phone.

I texted the KLM WhatsApp number and went back and forth with an assistant on my choices. Within minutes I was on the next flight, with the boarding pass on my phone. It was only later that I discovered that I had been dealing with next-generation artificial intelligence – in an example of the new field of conversational commerce.

If you haven’t encountered it yet, you will soon. Certain supermarkets are providing voice-enabled shopping services to customers, for example. In the US, Walmart shoppers can ask Google Assistant to add certain things to their virtual shopping trolleys and to learn from their shopping habits.

Google has similar deals with two other supermarket giants – Target in the US and Carrefour in France – while Amazon provides voice-enabled shopping in the UK to online customers of Ocado. Not to be outdone, Walmart recently bought conversation-commerce specialist Botmock to expand its services in this area.

There are already more than a billion people interacting with businesses via either text or voice-based conversational tools. In 2021, conversational commerce is expected to account for total sales of US$41 billion (£30 billion) worldwide, and is forecast to grow five-fold to nearly US$300 billion by 2025 – half of it from chatbots. So how is this market developing, and what does it mean for our shopping habits?

Coffee diehards and hyper-personal shopping

If conversational commerce still feels under the radar, one reason is that most growth has been in China, Japan and South Korea. All the same, it is cropping up everywhere. If you are talking to your girlfriend or boyfriend on Facebook and suddenly want to send them flowers, you don’t even have to break the conversation. You click on 1-800-Flowers.com, a conversational AI tool integrated with Messenger, and explain what you want. You don’t even need to enter card details if you use Apple, Samsung or Google Pay.

Or maybe like me you are a die-hard coffee lover. I used to stand in a queue to get my morning latte, but not now. I just order from my couch from the chatbot on the My Starbucks Barista app, and my coffee is waiting when I reach my local store.

The AI underpinning these advances encompass are deep learning, sophisticated natural language-processing, voice recognition, and cognitive computing – which is a system for machine-thinking that emulates human thought. But the big selling point – besides ease, comfort and shopping anywhere at any time – is probably the potential to make a customer’s retail experience much more personal.

If it lives up to expectations, customers might soon be able to interact with an AI who understands what they want in specific detail. We already see big retailers offering personalised products to attract customers – for example Nike and Adidas allowing people to design their own trainers.

But by using sophisticated AI, personalisation can move to a whole new level. Customers will receive personalised recommendations in their own language, easing the burden of choice and making the experience as enjoyable as possible. They might spend more money as a result – not because they are being manipulated, but because they almost feel like they are buying from a friend.

Meanwhile, businesses will gain new insights into people’s shopping behaviour. Yes this raises privacy questions, but it will also help businesses to refine their offering. This should reduce returns and increase sales.

Where it’s heading

Conversational commerce reminds me of the 2013 movie Her, set in a near future where Theodore (Joaquin Phoenix) falls in love with Samantha (Scarlet Johannson), an AI-based virtual assistant. The relationship eventually becomes unworkable when it emerges that Samantha is simultaneously having intimate friendships with thousands of men. She then combines with other AIs to perform an upgrade that leads to them withdrawing from human interaction.

We may be some way from falling in love with chatbots, but clearly there are questions about ethics here. The technology must not harm humans or pose any threat to their dignity. For instance, Microsoft recently restricted its voice mimicry technology because it makes it easier to create deep-fake videos.

Another issue is jobs. Automation is clearly a threat to the workforce, and conversational commerce could well be part of that. But unfortunately, businesses will not pay for so many support staff if AI can do the job at least as well. One consolation is that AI in its entirety might create more jobs than it destroys. For instance, the World Economic Forum predicted in 2018 that the net new jobs created by AI would be 58 million by 2022.

The day after tomorrow. Zapp2Photo

Looking further ahead, conversational commerce could become all the more prevalent in the metaverse, with voice-enabled shopping potentially accounting for 30% of all ecommerce revenues by 2030. It seems foreseeable that we will be interacting with AI avatars in virtual reality stores, or talking to bots in real-life supermarket aisles via augmented reality glasses.

What may seem alien to our generation is likely to be second nature to the shoppers of tomorrow. There are pros and cons to this technology, but I suspect my little chat with the KLM chatbot at Schipol airport will soon seem quaint compared to what comes next.