The Futurist and the Fellow: Retail

Geof Wheelwright: Welcome back to Arm Viewpoints. We have a bit of a special episode today that looks the future of the retail sector. Our guests today will be your guide to the retail experience of the future. With us are Rob Aiken, a fellow and director of technology at Arm and Matthew Griffin, CEO of the 311 Institute, in an episode we call the Futurist and the Fellow: Retail.

Rob is responsible for technology direction at Arm Research. He works on exciting things like distributed systems, technology, road mapping, and next generation memories. Matthew meanwhile is the founder and CEO of the World Futures Forum, and the 311 Institute, a global futures and deep futures consultancy working between the dates of 2020 to 2070, and is an award-winning futurist and the author of Codex of the Future series.

Matthew’s work involves being able to identify track and explain the impacts of hundreds of revolutionary emerging technologies on global culture, industry and society. Welcome to you both. I want to start with you Matt, to kind of kick off our conversation.

Where do you see the retail industry with things like shopping experience and logistics going in the next 10 to 15 years?

Matthew Griffin: Yeah, that’s a good question. So, uh, I suppose basically one of the first things we should sort of point out is that the shopping experience that retailers create really depends on the type of customer they’re trying to attract, the products that they’re selling and so on and so forth.

So for example, you know, if you have a look at a lot of the chatter around creating better customer experiences, if I’m buying toilet roll then I don’t care about customer experience. That’s it. Whereas if I’m buying a very expensive item, so you can typically think of it as cheap items I’m after I don’t mind a tactical experience, expensive items I kind of want that soft huggy, you know, customer service wraparound, basically that we sort of get when you sort of go to some of the actual premium retailers.

So on the one hand, we’ve got to have a look at the type of people that we’re trying to attract, the products that we’re trying to sell, and we create craft the customer experience and everything else basically around that. Um, second thing when we have a look at the future of retail, um, and we can dive into this in a, in a lot more detail, if you like.

Amazon is obviously the big elephant in the room. E-commerce is the big elephant in. But Amazon is now starting to lay out all of the different technologies and patents that it needs to become a fully autonomous retailer. So retailers, brick and mortar retailers think they’ve already been disrupted.

They haven’t seen anything yet. And then we can talk through the rise of virtual malls, which we’re seeing coming on, particularly strong in Asia. So particularly through the pandemic. There’s actually there’s loads going on and this is just kind of the snowflake on the tip of the iceberg. So to speak.

Rob Aitken: I guess I would, by and large agree with what Matt has been saying.

There’s been this bifurcation in experience of just absolute low cost, let’s get it done at the minimum margins and then the high end, I really need somebody to help me with this thing, but I think there’s a couple of other dimensions that have shown up and really been highlighted in the last year that, that are quite interesting.

So toilet roll was a good example, that there was quite an absence of that stuff, right at the beginning of the pandemic. And that really showed I think how fragile our supply chains are and that there are a lot of some assumptions baked in behind retail that turned out not to be true, like just-in-time sometimes becomes not at all.

And that, that in itself has all sorts of technological implications. But in addition to that from the actual retailer standpoint, There are a lot of questions around security theft and so on, which manifest themselves in different ways through online versus brick and mortar type stores. And so there’s, I think a lot of the technology is actually going to go into security and theft prevention in addition to improving the shopping experience.

Geof Wheelwright: Yeah. A lot of that would seem to tie back to logistics. So maybe we can talk a little bit about how you see logistics evolving and also scaling over the next few years?

Matthew Griffin: I’d actually team two things together. I would team fulfilment with logistics. So increasingly we’re seeing the rise of both fully autonomous logistics and fulfillment.

So I’ll give you a couple of examples of that. So in the UK, basically, we’ve got a grocer there called a Ocado. So even though they actually used to call themselves a grocery company, you go to their website, you buy food and it gets delivered sort of the next day, or whenever you like.

They are increasingly becoming a technology company that builds fully autonomous warehouses. And then they sell those to companies like Kroger so they can reduce their operating costs. You know, so we’ve on the one hand we’ve got that sort of autonomous fulfilment piece. We’ve also seen significant strides made in the future of artificial intelligence and machine vision, where we have robots that are now able to pick goods faster than human workers.

Now that’s significant, particularly when you actually have a look at the U S market as over 600,000 people work in the warehouse. So that has a variety of different impacts. When we have a look at logistics themselves, on the one hand, from a supply chain perspective, we’re seeing the rise of quantum supply chains.

So particularly with companies like BMW, where we are able to increasingly use incredibly powerful quantum computers to optimize global supply chains. We’re seeing supply chain transparency becoming increasingly important supply chain provenance becoming important as well. One of the biggest problems from a logistics perspective, that last mile we’ve got organizations like Mercedes, for example, who are building fully autonomous delivery vans with drones and robot dogs that literally spring out of them to do the last mile to your house.

And that’s before we start talking about the advent of drone deliveries which we’ve seen thanks to Amazon and 7-Eleven, one of the first drone deliveries that we saw was actually some donuts courtesy of 7-Eleven that didn’t come to my house. I was a little bit upset about that. So maybe I need to hack some of their drones and just send them off course, which is a security topic again.

But again, lots going on just in that space.

Rob Aitken: I don’t get donuts delivered, but I do get all kinds of other things delivered it and it, it is interesting. I think the last mile problem is going to be more challenging than people have thought because coming to my house is actually fairly straightforward, you drive down the street, you park, you go out the door in a way you are going to.

My son’s house is much more complicated. So where he lives the apartment building where he lives has a resident thief, whose job it is apparently to follow the Amazon truck and as soon as it arrives they steal whatever it was that was delivered and no one either at the apartment or at Amazon seems interested in fixing this so he and his fellow apartment dwellers have come up with all sorts of schemes and systems to get around this problem.

So that human ingenuity in terms of criminality is apparently almost limitless. But beyond that, just the technical difficulty of getting into some of these addresses, I think is going to prove quite challenging.

At some point, the technology will eventually be good enough to do this, but on the path from where we’re at to then there’s I think a bunch of breakeven points where it’s worth making certain class of deliveries in an automated way, and it’s just not worth bothering on some others.

So I think we’ll see this transition period where drivers of various sorts will still have jobs while they try to figure out how to actually make those kinds of deliveries.

This is a really interesting area because of the level of AI that’s necessary for some of these autonomous things. It’s again, it’s a classic example of the AI dictum that everything that you think is easy is hard and vice versa.

Geof Wheelwright: So as we kind of look out to this, what do we do with all this AI sensors and software to achieve a vision that everybody wants?

Matthew Griffin: So you know, when we have looked at say customer experience, one of the things that we’re sort of we can pull in is customer support. So for example, from a customer support perspective, you know, we’re increasingly used to the concept of I ordered something, I want to return something. I have a question therefore, I either go online, go to messenger, go to WhatsApp, go to, WeChat, go to whatever it happens to be.

But increasingly, retailers are able to use digital humans to service your inquiry. So rather than speaking to a text-based artificial intelligence bot, you can actually speak to a digital human like we see from companies like Soul Machines that literally looks like a human, responds like a human and will actually answer your questions and then start sorting things out for you in a very much a sort of similar vein basically to the way humans would.

But even when we start having a look at the rise of sort of fully autonomous systems, you sort of mentioned about the movement of technology basically into the landscape. I mentioned Amazon earlier. Now Amazon has recently used an artificial intelligence to design new fashion lines. The way it determines what good fashion looks like as it goes and has a look at all of the five-star rated fashion lines on Instagram, Facebook, Snapchat, all that kind of stuff, and it figures out this shirt has got a high rating, then that’s probably a good piece of fashion design that it should do something with, copy clone, ideate on top of.

You know once it’s designed things, we end up with sort of more intrusion into our personal lives because we have cameras now so for example, again, Amazon bought a company called Body Labs a couple of years ago. They will actually, you stand in front of these cameras, it will scan your body shape. It then creates a 3D avatar of you. So now it’s got your body measurements down to the millimeter level. So when you do go online and you start ordering that shirt that the AI designed. That shirt didn’t exist until that moment. So now we can use 3D printing for example, and Adidas and Nike are using this to eliminate inventory, which is another huge cost base of retailers. So now that shirt is manufactured on demand by a machine that just sucks in the load of fabric makes your shirt on demand.

It’s then fulfilled autonomously it’s then shipped autonomously and I get it but that shirt’s a perfect fit because that system has my body measurements.

Geof Wheelwright: All of this kind of points to an interesting evolution, particularly around the brick and mortar history of retail. Now you were talking Rob about being an expert in things that you didn’t expect to be an expert in. I never thought that I would have to worry about what does the phrase unexpected item in bagging area actually mean? And why should I care? But if I’m trying to get out of the store and that is preventing me from doing so I care. So those kinds of things I think are definitely impacting the retail experience with technology. I’m wondering how you see further change on the brick and mortar end?

Rob Aitken: Automated supply chains present all sorts of opportunities for counterfeiting as well and so provenance as Matt was mentioning earlier is really important. And I think we’ve been doing a bunch of work into how you can actually for chips, get them so that they’re able to identify that this is in fact, the chip that you have is the one that you think you have, and it has been registered with some authority, some place that vouches for its correctness and things like that. I think that there are various technologies available that will allow for that to expand to things beyond chips.

So for example, the little Wi-Fi tag or the RF tags that you use to identify merchandise. You can start putting things like that into the labels. And those can start identifying that this is in fact what you want. I mean, shoes is a classic example. I ordered a replica. So I had a pair of shoes and I liked this pair of shoes so I went online and ordered another version. You know, I want another one of the same shoe. And the thing that arrived, looked like the shoe that I had before, but it was actually a counterfeit which I could tell because every time I walked it went squeak, squeak, squeak on the floor.

It was fascinating if you looked at the online reviews, the reviews were like, yeah, it turns out that this is actually a counterfeit because it goes, squeak, squeak, squeak when you walk on the floor. And so apparently other people had had the same experience and some of what Matt is talking about, if you can get the production closer to the person, so it’s made on-demand and we know who’s making it and who shipped it and so on. This kind of automation and guaranteeing that what you get is actually what you thought you get and that no-one’s intercepting it. There’s a lot of opportunities there for establishing trust in these supply chains that in many cases can be automated with suitable tracking technology, which then leads into its own question of, okay, who is doing this tracking and what are they doing with all that information that they now have.

But that’s yet another discussion but there’s definitely a need in an autonomous system to keep track of a variety of low-scale and high-scale fraudulent efforts to corrupt it.

Matthew Griffin: Well, I mean, you know, so I mean, you know, Rob raises loads of lots of very good points. One of the things I’d sort of put in here is, um, yeah, I ordered, I ordered a pair of replacement jeans.

They pick out jeans, put those into a fulfilment box, but that fulfilment box reads that RFID tag and it says this is authentic. You know, we’ve got loads of different ways to prove the authenticity of goods. Nanomaterials nanotechnology, nanosensors, nanoantennas, you know, I’ve got a big list.

Now if for example the system knew I ordered some jeans and then picked up some counterfeit jeans that didn’t have that technology actually built into it. Then there should be an alarm that’s raised that sound’s going, hang on, these can’t be jeans because there’s no RFID tag or the RFID tag or the crypto anchor. Uh, for example, IBM have got one of those.

If the crypto anchor is wrong therefore these might be counterfeit. Let’s go and send those off to go and be investigated. Meanwhile, let’s go and pick these from another part of the warehouse or whatever it happens to be. We can do this. It’s kind of like Amazon Go where you can walk into an Amazon Go grocery store, scan in.

You don’t have to pay on the way out because the store has scanned everything. It’s scanned all the products, it’s scanned what you’ve picked up and put in your basket. It’s calculated the prices. Why can’t we put those kinds of technologies into modern fulfilment warehouse. The technology is there. It’s straightforward.

The cost of putting these little nanoantenna into products is like a millionth of a cent. You know, it’s incredibly small and yet that would help alleviate – you’re never going to get rid of fraud, but that would help eliminate quite a lot of fraud, counterfeiting and everything else. And just on the fraud numbers. The serious organized crime industry is growing at about 19% per year, because it’s something I track. It’s now valued at $6 trillion. It’s the fastest growing industry of every industry. And that’s before we talk ransomware and all those other things that they’re making money out of but counterfeiting is a huge industry in itself and there are solutions.

Geof Wheelwright: Yeah. And it seems that there is a piece around ensuring trusted and secure experiences in retail, as you were just discussing with fraud. Maybe Rob you can start us off on that?

Rob Aitken: There’s one of the issues that you brought up is, is the individual tracking of objects and building an object that can be individually tracked depending on the technology is not that hard.

There’s these physically unclonable functions, for example, that you can build there’s various non-volatile memories that exist and can be made to exist. I think the biggest challenge is actually building the logistics behind that. So building the attestation service, that’s able to identify that you have, in fact, this number that associated with this particular tag is in fact, a legitimate number for your pair of jeans.

And there’s then where is that tracked? is that tracked by the jeans manufacturer, by the retailer, or by somebody else? And depending on what answer you provide to that question, then that brings up a series of costs and complexities associated with keeping track of that database. So there’s I think the cost is, is mainly in the service part of it, rather than in the actual tag itself, which as you say can be pretty cheap.

Matthew Griffin: Well, I mean, and you’re right. You know, if you have a look at how you can actually track and tag these. So on the one hand we can use sensors that are actually within a warehouse, but you know the vendors and the distribution partners agree who does what you know. So that’s a sort of demarcation of roles, if you will.

But just sort of giving you an example, Rice University last year developed a 3D printable graphite tag. That you could eat. So it all, it is, it just looked like a very tiny, straight little black straight line, and this can be printed on things like strawberries. So this doesn’t just apply to jeans. You know, in this particular case, you can 3D print a tag that will transmit a signal.

And an authenticator to a local machine saying this strawberry is authentic. And then it also tells you everything about that strawberry that you want because it’s linked to blockchain provenance and that kind of stuff. But you’re right. You know, when you have a look at the implementation base, you have the systems to support these types of technologies.

That’s where it comes down to an ICT business case. And this is where it says we’re losing $300 million in fraud. If you look at the US there’s over something like $250 billion of medical and pharmaceutical related fraud. So you look at that part and you go, it’s costly. Fraud is costing us $300 million, this solution’s going to cost us $5 million to implement, what’s the business benefit etc and you sort of run it as a traditional program but for all of them.

But for everyone who complains about counterfeit goods, Congress complains about counterfeit goods by sea on the big e-commerce sites. The e-commerce sites say that they are continually monitoring the situation. There’s one giant that’s spending billions of dollars trying to identify counterfeit goods. These projects, though, these types of ICT projects are not going to cost billions of dollars to implement and they’re going to get rid of some of the headache, maybe not all of it, but nevertheless, we can still move the dial.

But instead as consumers we are still going on to a lot of these different sites, crossing our fingers, going “I hope this isn’t a counterfeit thing” and believe me, I’ve ordered loads of counterfeit things. Well, my mainly my kids have really, to be honest, not me obviously. Um, yeah, we’ve had loads of counterfeit things coming through because the fraudsters are very sophisticated in the way that they wrap everything up and they package the reviews and the way that they scam the system, but this stuff gets rid of that.

Geof Wheelwright: I think you’ve given a great overview of the kinds of things that we’re going to need to do to make things more trusted and secure. Maybe we can kind of pull the lens back a little bit more and kind of look more broadly at if you look out over the next 10 to 15 years, what one thing excites you most about the future world of retail and why?

Rob Aitken: We’ve touched on this briefly, but, but one of the things that has popped up in e-commerce so far, is that there’s now access to a huge range of things that we didn’t have access to before. It used to be that if I wanted say a particular record album or whatever, I had to figure out which obscure store in which obscure town that I was visiting, had a chance of stocking this thing and then make a special trip to go there.

And it was all very exciting, but now I can just, you know, any band that’s ever recorded, any song anywhere I can just go click and there now I have it. So there’s that I think extrapolating that forward into a world where you can get anything, but you can also get a customized version of anything so that I don’t have to worry about whether or not this shirt will fit, but that kind of future to me is really exciting.

And I think that the opportunities for uniquely tailored products that match our tastes perfectly is to me the most exciting thing going forward.

Matthew Griffin: So I suppose, yeah, as an avid shopper, not, you know. I’m one of those guys that gets dragged around shopping centers, frankly, along with my kids.

So I suppose my wants basically would be, I would love to be able to order something online and get what I actually expect. So for example, shirts that fit, furniture that isn’t half the size of what it looks like in the photo. That’s a fun one. Um, you know, we tried sending it sideboard back, done that during lockdown. That was insane.

But then sort of on a more pragmatic point. Yeah. The ability to use, for example, on demand manufacturing technologies to personalize and tailor goods in the way that I like. That’s exciting to me. And that’s again, where we start pulling in things like, you know, hyper-personalization advanced manufacturing and all kinds of other things.

And from a customer experience perspective, you know, surprise me by being good. We talk about retail being sort of an experience, but I guess, you know, if we sort of ask the audience here, how many of them have a good experience with retail? Most of them would probably go “it’s not bad, it’s not good. It’s a bit in the middle.” Is that really the experience that retailer, the vast majority of retailers are going for? Because that’s actually what most of them are delivering at the minute. So I would just as I say, I would like to be able to order stuff and get what I actually expect. And then if it can be personalized and everything else, basically, that’s just a sort of cherry on the top.

Geof Wheelwright: Well, I’m sold, which is really what you want from a retail. Thanks to you, both Matt and Rob. Thanks to everyone who joined us to listen to, into our podcast. Today, we look forward to bringing you more conversation in the next episode of armed viewpoints. Thanks again for listening today.