putilov_denis - stock.adobe.com
The role of AI as an everyday life assistant
We speak to the co-author of a book that explores the idea of artificial intelligence-powered automation that enables machines to become customers
In 2019, Amazon discontinued its physical Dash button, a gadget that allowed customers to reorder products at the push of a button. But this idea of automatic replenishment is set to make a comeback – except, according to Gartner fellow Mark Raskino, it is likely a machine will be doing the ordering and replenishment automatically on behalf of the customer.
Raskino is co-author of When machines become customers. He recently spoke to Computer Weekly for a podcast discussing the idea of machine customers, which he describes as “something that shops for you or, eventually, will shop for itself”.
This is not hard to imagine, and there are a number of examples – post Dash button – that show how quickly the technology is being accepted. Amazon itself has a range of methods of easy purchasing, such as via Alexa, subscription services, or even digital dash buttons and smart shelves.
Raskino says there are three phases of evolution. The situation today, which he describes as “base zero”, is where devices alert the user that they need replenishment, such as the light on a dishwasher that comes on when the salt needs topping up.
Phases of machine customer
In the book, phase one is where a product has a tight binding with a particular e-commerce activity. Automatic printer ink cartridge replenishment is an example of a phase one machine customer. According to documentation for the Amazon Dash application programming interface (API), Alexa can be used to keep track of usage and automatically reorder when the ink supply is low, based on alerts from a usage sensor.
Similarly, a coffee machine can be programmed to order pods automatically from the manufacturer when the supply goes below a certain threshold. This simply requires an internet of things (IoT) device, fitted with the correct sensors, that is capable of calling an API, such as the back-end service behind Amazon Dash.
Raskino says Illy has moved on to the stage where the machine truly becomes a customer, by closing the replenishment loop via the Amazon Dash Replenishment Service. Illy’s espresso and coffee system integrates the Dash service to track coffee capsule usage and automatically reorder more.
Raskino describes phase two as the point when the machine has a level of intelligence, which makes it an adaptable customer, with the ability to interact across multiple services and providers.
Today, he says, it is entirely possible to ask a smart speaker like Amazon’s Alexa for a recipe and have the AI find the recipe and ingredients, and even purchase them from Amazon. This, he says, is about delegating the work of being a customer to that machine. The machine has found the recipe, decided what ingredients are needed and then shopped for those ingredients.
One may question whether the AI is working in the customer’s best interest, since it will order items directly from Amazon, but Raskino believes this is no different to the situation where a child is asked to go to the shop to buy potatoes – the child may choose cheaper potatoes in order to have some spare cash for sweets.
In phase two, Raskino says the adaptable machine customer starts to infer a need on behalf of its owner and will then be able to shop for these items from different sources, maybe choosing those that are better value or the healthier option.
Looking back at the recipe example, Raskino says: “The machine might be providing healthy choices because it knows you’ve been drinking quite a lot of alcohol lately and maybe you need more vitamins.”
This is not a simple replenishment action since the machine is making a more informed decision based on an overall understanding of the machine’s human owner. Similarly, a smart vacuum cleaner may be able to sense that the owner’s pet cat is losing more hair than normal so it may automatically order supplements or book a veterinary appointment.
Raskino says surveys have found that people whose working week has been reduced from five to four days spend their free day doing chores. “Shopping is a chore. We’re not very good at it and we don’t have the time, so it is easy to see how a machine could do a better job,” he adds.
The same is true of shopping for the best energy provider, or car or home insurance. Raskino says people are already accustomed to going to a price comparison site to buy such things. They also understand that the comparison site probably takes a small commission and may not offer absolute value for money over manually searching for the best deal, but it saves people a lot of time. Raskino sees no reason why a human needs to be involved in shopping for such services at all. An AI could be fired up annually or when the service is up for renewal to search for the best new deal.
Read more about automating business processes
- We speak to chief procurement officers about the role of artificial intelligence in helping their organisations reduce costs and tackle the economic slowdown.
- Automation and AI is improving how the service delivery team at Leeds Building Society assesses the business impact of an IT service degradation.
This idea of the machine coordinating tasks, purchasing and ordering services, fits in with people’s busy lives and, according to Raskino, integrates well with today’s gig economy. In fact, he sees plenty of opportunities for people to work in an ecosystem where the machine is a customer that has the ability to coordinate activity with human workers.
For instance, a smart oven may sense it needs a clean and book in an oven cleaning service itself. In this scenario, a home equipped with smart door locks would be able to send the oven cleaner a one-time PIN for entry to the house.
It is not just the largest e-commerce platforms that will pick up business from these machine customers. Raskino says platforms Etsy and Shopify offer e-commerce for sole traders and small businesses.
“There’s a multiplicity of software providers that work for very small businesses. I’m always amazed at the number of specialty software providers that have popped up, such as businesses that offer card payment terminals through to those providing electronic shop signage.”
He believes there are also opportunities for machine customers to help small business owners, and this is one of the areas where phase three machine customers can make a difference. “Small business people are up to their armpits in admin and trying to get the best price for the items they’re going to resell or for the parts they need to make a product,” says Raskino.
One example given in When machines become customers is a smart fizzy drinks dispenser in convenience stores. As Raskino points out, figuring out what fizzy drinks are selling best is a quite a dull, repetitive task for shop owners, who are also visited by salespeople from drinks distributors trying to encourage them to change the mix of drinks in their fridges.
“The whole thing is quite inefficient,” he says. “We’ve been talking to a couple of drinks distributors who are working on smart fridges.” Such a fridge would have access to local weather conditions and understand the make-up of the people visiting the convenience store, along with knowing what drinks are being picked up. In effect, the smart drinks dispenser has become the customer and autonomously replenishes by predicting demand.
One of the concerns the book raises is how businesses experienced in selling to humans will respond. There is no reason to assume the machine will remain in the domain of low-value purchasing, leaving businesses free to focus their efforts on high-value human customers.
“Doubling down on the human market and perceived higher-value human customer service capabilities, the losers will find their cost of sale gradually increasing even as their revenue and total addressable market appears to shrink,” warn Raskino and co-author Don Scheibenreif.
Society may not yet be ready for the machine customer, but the idea is finding its way into people’s lives by automating boring or repetitive tasks.
In the book, Raskino and Scheibenreif discuss the May 2018 demonstration by Google CEO Sundar Pichai of an AI assistant called Duplex. The AI was so convincing that it was able to book an appointment at a hair salon over the telephone, without the person on the other end of the line being aware that it was a machine making the appointment.
Raskino and Scheibenreif believe Duplex was ahead of its time. The demonstration caused a public uproar and was quietly put to one side – but five years on, the ongoing development of generative AI shows just how convincing a chatbot can be at holding a conversation with a human.
Much of the technology already exists today. It is only a matter of time before such advances in AI are integrated into people’s lives, to automate chores, order services and purchase products on their behalf.