Beyond the Checkout: How AI Is Redefining Retail from the Inside Out | Special Episode | Learning Kerv | Learning Kerv

Description

⭐ Special Episode! ⭐ Beyond the Checkout: How AI Is Redefining Retail from the Inside Out

In this special episode of Learning Kerv, host Rufus Grig is joined by three retail experts to unpack how AI is really changing the game in the retail world. From transforming customer experience to reshaping in-store operations and supporting frontline workers, this episode gets into what’s working, what’s not, and what’s coming next.

Guests include:

Neil Holden (Ignite AI Partners): on why AI is the biggest shift retail has seen (yes, bigger than dot-com).
Dan Speed (Dunelm): on why your fancy AI is useless without rock-solid networking.
Tim Walpole (Retail Trust): on the link between AI, wellbeing, and your store's bottom line.

Expect straight talk on:

Real-world AI use cases in retail
Data ethics and the customer trust contract
AI-powered colleague support and wellbeing
Why infrastructure is the unsung hero of innovation
The not-so-distant future of voice commerce and AI assistants

If you want honest insight (with a side of wit) on what AI means for retail right now - have a listen!

Hosted on Ausha. See ausha.co/privacy-policy for more information.

Transcription

Rufus Grig
Hello and welcome to The Learning Curve, the podcast from Curve that delves into the latest developments in information technology and explores how organisations can put them to work for the good of their people, their customers, society and the planet. In our first series we had a look at all things AI and in particular generative AI and in this special edition of the podcast we're going to look very specifically at some of the ways that AI is reshaping retail. not just for customers and for customer experience, but also for a huge part of the workforce, the UK workforce that works in retail, and also for culture and sustainability in retail too. My name is Rufus Grigg, and today I'm joined by three very special guests, all from the retail sector themselves, who came together a couple of days before we're recording this podcast to lead a seminar on the use of AI in London. And they've very kindly agreed to come back and share some of their thoughts and insights with a wider audience on the learning curve. And I'm particularly grateful as I wasn't able to attend the event. So I'm actually hearing this for the first time. So let's meet the guests. First up is Neil Holden. Neil is a veteran, a very youthful veteran, I would say, of the world of retail technology, now offering the benefits of experience in a consulting capacity at Ignite AI Partners. Welcome, Neil.
Neil Holden
Hi, Rufus. Thanks for having me.
Rufus Grig
You've seen a lot of change in technology and retail in your time. Where would you class the current position of AI amongst those sort of major changes that you've witnessed?
Neil Holden
Blimey, yeah. There's definitely been quite a few. I mean, when I first started in retail, it was green screens and computers had not long been introduced into the workplace. And there's been many variations of changes since then. Dot com, online payments and so on. But I can't stress enough how I think... this latest revolution in technology is the most transformative. In fact, it's why I've given up a 27-year career in retail to go and get after it and hopefully help other businesses gain some advantages from it as well.
Rufus Grig
Well, it'll be great to hear some of your thoughts on that when we get stuck into it. Secondly, we have Dan Speed, Dan's networking technology lead at Dunelm, which is one of the largest homeware retailers in the UK and still growing at a really impressive rate. Great to have you with us, Dan.
Dan Speed
Hi, Rufus.
Rufus Grig
Dan, just for a moment. imagine there's someone listening to this podcast probably overseas who doesn't know Dunelm. Can you just give us a quick introduction to the business?
Dan Speed
Yeah, so Dunelm's a omnichannel retailer. We've got a store presence of 200 and growing and we also have a large online retail presence as well. We specialise predominantly in the homewares market so everything you can think of for your home you can get at Dunelm.
Rufus Grig
Brilliant, thank you Dan. Really good to have you with us and then our final guest brings a slightly different perspective. The perspective of the retail workforce. Tim Woolpole is Head of Product and Innovation at the Retail Trust. Tim, great to have you with us on the podcast.
Tim Walpole
Thanks for having me. Great to be here.
Rufus Grig
And for those who don't know the Retail Trust, could you just give us a brief overview of what you do and your mission?
Tim Walpole
I guess in summary, set up in 1832 to look after the well-being of people in the retail industry. And since then, the ethos hasn't changed. We're here to support. The well-being of colleagues in the retail industry. We support about 650,000 colleagues for about 220 retailers.
Rufus Grig
Brilliant. Well, thank you very much for joining. It'll be, I think, absolutely vital to hear the perspective of colleagues in this transformation as well as all the other stakeholders. If I can turn to you first, Neil, with your sort of broader perspective, you know, what sort of things should retailers be thinking about with AI in 2025?
Neil Holden
I think 2025 represents quite a shift, really. So... 2024, early 2024, was when I established the business Ignite AI Partners. And I think we were just entering into a market which was quite chaotic. In fact, I read a recent report which was explaining the dynamics of the AI industry from 2024 to 2025. And 2024 was deemed to be quite an explosion of different models and solutions and a very chaotic sort of landscape. I think 2025 represents. a bit of solidifying of the foundations of the AI industry. So models are maturing, actual results are being seen to be driven from the solutions. So I think from a retailer's perspective, I would have anticipated that last year I'd have been very confused. I think they probably still are quite confused, but I think it would be the uber confusion in 2024 about what's good, what's not. Coming from the retail industry, particularly working for businesses that are long established and have a heritage, De Devoe. jumping into the newest solutions that are just into the market they want to see that it's proven that it gives results that there's less risk around that as well in terms of that adoption ai is changing the game i think where that's concerned because a lot of retailers are going into things quite early and testing and trying to get results from them but even those that are a bit more risk averse. I think 2025 will be more. about, okay, we understand this a bit more now. We're probably working with partners who are helping us to understand that as well. Let's actually do some real implementations using AI.
Rufus Grig
Yeah. Okay. Interesting. So, I mean, I think the timing piece is interesting. I mean, if we think about the sorts of challenges that retailers are faced, you know, changing consumer buying habits, you know, movement away from high street, retailers' perennial challenges with margins and supply chain and everything else. Now, what sort of issues that retailers face are being addressed or are addressable by AI? Where are those applications?
Neil Holden
Yeah, well, there are many. It is vast. You know, we talked in the session about how AI could be seen as more of a utility like electricity than a bit of software, which is great in one sense, scary in another, a bit confusing in another. The real key, though, is to understand where those problems can be addressed using AI-related technologies. And if you just say it all can. that's not very helpful. You really want to pinpoint exactly where the biggest benefits are, where the biggest problems are that can be solved. Retail, as you just mentioned, you know, has faced many headwinds over the last five, 10, even longer years. You know, in more recent years, it's been about costs, you know, inflationary headwinds, consumer confidence, as you said, general cost of living crisis, and all of those things hit retail quite hard. Where AI can lend a practical hand to those things is around efficiencies. So things like intelligent automations, helping businesses to do things quicker and more effectively using AI-based technology. Also understanding your business and your customers in far greater depth than ever before and more quickly by harnessing the power of data with AI technology combined either within that or on the front end. So we see lots of our clients asking for things like. natural language interfaces into their data. I want to talk to my data. I want to understand things in natural language. And that solves a couple of problems. Firstly, there's less of hard data mining and data science required because a lot of the solutions can now do that. And so it democratizes that for people who aren't data scientists. I can still get great insights from data, but use AI technologies to achieve it. And secondly, to understand consumers at a much greater depth than ever before. Things like sentiment, you know. Call centers have long had analytics that say, this is how many calls you've had and what they were about and what you should maybe do next. But now AI brings things like sentiment analysis to the table far more quickly and effectively. So you can not only understand all of that, but you can also understand, how do my customers really feel about us as a business? How was their last interaction? Were they happy? Were they annoyed? Were they angry? Is there something that we need to do there to then follow up and recover that conversation? the vastness and the depth of capability that... that AI can provide for retailers is quite extensive.
Rufus Grig
Yeah, I think it's similar sorts of challenges that the businesses outside retail would be looking to AI to solve, I guess. I mean, one of the things we've talked about in other editions of the podcast is that the state of AI now, particularly things like generative AI and cloud-delivered AI, is that it is having a sort of democratizing effect in that where it was the preserve of the very largest of organizations that could afford the compute power. the the data scientists just the implementation costs it's now you know within the realms of much smaller organizations is that something that's happening in retail are you seeing a leveling or is there still a divide between those who can afford to use this sort of technology and those who can't there
Neil Holden
is still a divide for sure there's no getting away from that i mean the democratization of ai technology is nothing like i've ever seen before you know i can sit here in my dining room and get access to some very powerful technologies as an individual, not as an organization, and do some quite clever things with those. So if you then factor that into a smaller business, you can do similar things. But the divide is where doing proof of concepts or small projects using open source or free tools is one thing. But if you then want to scale that across your business to get real advantages from that over a longer time period, that's where the cost divide does come in. The leaders in retail, people like Curry's, who have really adopted AI technologies to a big degree and are seeing great successes from that, have an alliance with Microsoft and no idea what the budgets are, but I imagine they're very large. For a smaller business to do that to that degree would be cost prohibitive at the moment. However, the cost profiles have to change with the underlying technologies, and I think they will. I think it's inevitable. but at the moment costs can't continue to increase in line with the capabilities in a direct correlation. There will need to be a leveling off of that. And hopefully then that will give more advantage to small and medium-sized businesses to still get access to these great technologies, but not with exponential costs.
Rufus Grig
Yeah, okay. I mean, you talked a lot about data and the use of data and the ability to analyze that. We think about sort of the responsibility in AI and openness and explainability too. what's the sort of Contract with customers around the use of their data, your retail organization's use of their data. You know, how open are organizations being on that? And is it something that customers and consumers are going to be happy with?
Neil Holden
Yeah, that's a good question. So I think the industry is still figuring this out. Legislation and regulations are catching up, but not quick enough. Common sense prevails in most cases that businesses understand their consumers' concerns and requirements about how their data is used. I mean, that's not a new thing in GDPR and the right regulations are in place for that. What AI does is makes that 100 times more important because it puts a microscope on the potential misuse of data and how quickly that can turn into an actual issue because of AI is just ever so heightened. So businesses, I think, are trying to understand, well, we're doing all these things already in terms of data protection, information security, and ethical use of some of these tools. Is that correct? Is it not? How much do we need to tell people? Our advice to businesses is you can't be transparent enough. And I think if you are, and you're... behaving responsibly and ethically and and openly in terms of that and telling your customers where you're using ai where you're not how their data is stored allowing them to opt in in the correct ways then you shouldn't go wrong where that's concerned but we're often asked what new legislations and bits of regulation do we need now need to comply with and unfortunately i mean the eu ai act is the latest most up-to-date piece of legislation but it's I'd say it still has a load of gaps and it's not even enforceable until 2026, which I find a bit baffling.
Rufus Grig
Yeah, it's interesting, isn't it? The pace of technology development is outpacing the ability of legislators and regulators to remotely catch up and get ahead, which is a slightly troubling state of affairs. Thank you. If I could turn to Dan, I know that you were speaking when we were together at the event about how do we actually... place the groundwork to enable some of this technology to be able to be in use you know throughout the retail estate and we think about you know computing as being the important part of AI and the big number crunching and the GPUs and the data centers but actually if you can't connect if you can't actually access that compute or you don't get that connectivity in store none of this stuff is going to work so what's sort of the big problems that you were trying to to solve yeah so obviously I've got a slight bias in this area but for me networking is
Dan Speed
your absolute foundation to any of this from both sides so obviously you've got the networking that underpins the compute that generates and was working on the ai but equally if not more important is the networking in the area of where you're consuming it so for us in our store estate the foundation has to be there in you in your retail network you have to put the groundwork in and work with a good wi-fi provider to make sure you can put that in securely and to To the point around the experimentation with AI and within retail, there's a lot out there, there's a lot of tooling that retailers want to get their hand on and to start and see what that does. But there's obviously questions around new hardware coming into the estate and new devices that we want to be able to put in and we want to make that available and see what that can do for us. But obviously there's concerns around putting devices on the network that might have access to things they shouldn't so that security element it's foundational to being able to do that without having the platforms in sort of what we've done with Juniper most there's no way we'd be able to be in the position I if I think to where we were five years ago with the the hardware we had in the estate to be able to turn around to a one of the product needs and say yeah that's fine you can connect that device here here's a key and off you go we definitely wouldn't have been able to do that whereas now those conversations come and go a lot quicker let's just wind back a little bit you mentioned juniper um so you've
Rufus Grig
put this Juniper Mist Wi-Fi into the store estate. Tell us what's different and special about that and then perhaps it'd be really good to unpack the idea of some of the devices that you are wanting to connect in because you know for a lame person like me you know in-store Wi-Fi you know I tend to think of that as a thing that a customer connects to so that they're on their phone rather than something that's being used actively by the retailer to improve my customer experience?
Dan Speed
Yeah, so... The idea of being able to have those devices in the store, so we've looked at certain cameras, we've looked at devices that count customers coming in and out of the store, we could potentially be looking at systems that, as Neil mentioned, are gauging the mood of the customer, and we can be looking at services that support new technologies going in and around self-checkout. They're all devices that need something to be able to connect to the internet. most of the AI processing that's done in the cloud or it's done a service remote from the device because the amount of power that's needed to do that and to be able for that to happen you need to know that your device has got a constant feed to what it's trying to connect to so MIS has this idea of service level experience and that's really well tracked through the platform and it's using its AI and machine learning in the background to process what's going on with that and to be able to give us a very quick scorecard of what's going on with the estate.
Rufus Grig
So you're using AI to deliver the infrastructure over which you can then operate more AI. So eating ourselves here. That's really interesting. And I think I'm totally with you that that infrastructure, that connectivity layer and the reliability of it and your ability to just to be able to deploy things and innovate and experiment. And the answer that you can give back to the business more often is yes, which is, I guess, what you always want to be able to do, isn't it?
Dan Speed
Yeah, 100 percent. We I remember being in a world of networking where the most common response I gave to a question was no, or it's going to take a while. Now we're going from a point of yes in sectors where we've not had this level of investment and we've not had these excellent products we've got now. You've almost wanted to shy away from conversations because you know ultimately it's going to get to a no or it's a lot of work or you're going to be giving out disappointment. It's no longer a utility, it's a product. It's part of that journey of bringing something to life in our stores.
Rufus Grig
No, that's brilliant. Thanks very much, Dan. Tim, if I can turn to you for a minute. Dan mentioned what the Wi-Fi has enabled them to be able to do in terms of being able to equip colleagues in store with useful tooling. I think I'm probably guilty, as maybe some others, of not really grasping the vital importance of retail workers on all of our lives. You know, and it was the pandemic that really kind of brought into critical focus the fact that, you know, they're real frontline workers sort of exposed to the public all day, every day. Could you elaborate a little bit on how colleague well-being can sort of impact upon retail experience, on the performance of the retail business, the performance of the store, as well as on clearly, most importantly, their own lives and family situations too?
Tim Walpole
Yeah, I think it's a really interesting question that is probably not thought about. So if I take, I guess, an example of maybe someone that's a retail colleague and Someone has come into that store and either physically or verbally abused that person, which as we know happens on a daily basis to most colleagues. They go away and it's impacted their mental health, for instance, and they come to work to the next day and they really don't want to interact with customers because they've had the experience the day before, they're not feeling quite in the right state of mind. So they come to work not mentally prepared. And the effect on that is that the next customer comes in and they don't want to deal with them and they don't go out and say, hey, how can I help you today or something? It's just not I'm going to sit here. What does that do, of course, to the experience for the customer? And this negative experience, I walked into the shop, no one helped me. And how does that go? Well, that affects the store performance. It affects the financial performance. So you can already start seeing the experience from that customer, that mental health support or something just because of that one item of abuse. really impacts not only the colleague, but therefore the customer, and therefore the business as a whole. And at the Trust, we realise that's important. And we want to be able to say, how do we support you with anything that may have happened? So how do we understand, maybe through Juniper, maybe through understanding that a colleague has been abused, that we can reach out to them, we can support them with some proactive content to be able to say, here's something to support you, such that when you come to work the next day, you're going, well I've had the support I need and I can I can cope with that with that proactive support and that help.
Rufus Grig
So are you talking about colleagues use of things like the handheld devices that Dan talked about to be able to report you know instances of abuse or instances of challenge in in the workplace and are you able to use those same sort of devices to be able to you talked about pushing content to be able to perhaps support people in in that instance?
Tim Walpole
Yeah indeed so whether it's those handheld devices or some other signal that allows us to be able to identify that. a particular colleague has experienced abuse, but maybe also which other colleagues were near that may have also seen that abuse and therefore is potentially affected. And yes, we have a large amount of data in the retail trust. As I said, we've got about 650,000 colleagues on the system. We have a large amount of preventative content to be able to say, actually, we could then reach out to you automatically. So just by identifying that person, we can then reach out automatically saying by the way, do you know the trust is here to support you? Here's some proactive content, or potentially maybe you need some counselling, some in the moment support or something. So by having that signal from that retailer back into the trust, we can then reach out to them to be able to say, here's the content, here's how we can support you. I mean, I guess the other part of it is we can actually therefore measure the impact that this has, and we can show the impact that this has. At a couple of levels we look at absenteeism. We look at presenteeism, which is coming to work and not being in the right state of mind. And also Collie Churn and all these things have a major effect in the industry. But by bringing this content in to support them, we can show the impact that this has. back to that retailer, not only in financial performance, but also their own well-being. And the bigger impact is social and economic impact. We can show there is a, through happier colleagues, there is less absenteeism, happier customers, which actually leads to a positive impact to the GDP of the UK as well.
Rufus Grig
Yeah, clearly, absolutely vital. How do you deliver content? I mean, we tend to think of, if I picture a retail worker, they're in situ, They're on the shop floor. not sat at a computer terminal. They're also typically very busy responding all the time. How do you actually get content to people? How do you sort of schedule time for them to be able to have that support that they need in those environments?
Tim Walpole
Yeah, that's a really good question. And actually, we work as a B2B2C. So we actually send our content to the retailer, because the retailer, of course, has the ability to push that content out. But actually what we do is through the use of really clever technology, through the use of AI, we can understand all the colleagues in a retailer. So we can understand potentially that some colleagues in the north of England are struggling with abuse or financial support. And we can tailor campaigns back to the retailers. So we're just building out the latest version of our platform using agentic AI, one of the latest buzzwords. But what that does is it allows us to write campaigns on the behalf of a retailer, in their brand, in their tone of voice, because of course retailers are also time poor in being able to send some of this content out. So through the use of AI we write that campaign for them. So in this case here's a campaign to support colleagues in the north of England struggling with abuse or financial support. We will write that on their behalf and they will send it out to all their colleagues who they have the contact with. But we've identified already who the colleagues are, what they should do. And we then track that on the retailer's behalf saying, by the way, do you realize this is the impact that this campaign around this has? And this is a financial and economic impact that it's had just by doing that campaign.
Rufus Grig
So the AI is enabling you to, I guess, to turn these campaigns around more quickly, which means you can support more people, more retailers, more retail employees. Have you got any sort of idea of just how much the... You can quantify how much the move to using AI for this is extending your reach over and beyond using traditional tooling.
Tim Walpole
Indeed. So we used to write all campaigns by hand in the retail trust brand, of course. And then each retailer had to take that and turn it into their own brand. The challenge is, of course, if I was to do that for 220 retailers and I was to do three campaigns a month and I was to do it in their tone of voice, the amount of... staff that we would have to be able to do that is just non-justifiable so not only as you say is it allowing us to reach the correct colleagues with the right content in the right tone of voice every month or potentially twice a month targeted to what's important but it also says we can do in a way and as you said that is hits the right content and we've already shown through our trials that there's been a 50% uplift in people reading content through the retail trust, through targeting the right people, automatically in the right tone of voice. We're here to support more colleagues. We're here to support the better retail industry. We're improving the well-being. We're reducing absenteeism, presenteeism term, which is a good thing for people. Business in the wider economy, it's great for the retail industry as well.
Rufus Grig
Yeah. I mean, that's an amazing story of how the technology is helping, but also such a win-win in terms of the positive reinforcement cycle of being able to react and support people in those situations and the ultimately financial outcome as well as welfare outcome you get at the end. Was there any challenges you had, and I'm going to come to the others on this as well, around training or upskilling people to be able to adapt to using? I'm more thinking of your immediate colleagues in transitioning to using these AI tools.
Tim Walpole
Yeah, it's a really good question. And it's easy just to throw AI out and say it's going to do write stuff but it's really important to do it ethically and responsibly. making sure you're using the tooling correctly. And as part of that, our first insight is we need to be able to bring the business on board and we need to have the governance and policies in place. So the first thing that went in is that here is our ethical and responsible AI policy that says how we're going to use AI. Here is some training for everyone to understand what that policy is and what it means, the EU Act, et cetera, and as they change, being able to bring people on that journey. but also making sure that the content that we produce is actually suitable to be able to send out. So our policy says there's a human in the loop in every single conversation we send out. My normal example, and I talked about it the other day, we had a piece of content where it suggested that actually all the retail colleagues should just take CBD oil. It's probably not the right thing to do. The human in the loop picked it up. Yes, and it's really important to make sure that you use AI in a responsible way and you have a human in the loop. We need to make sure our brand is truthful and honest and trusted, and that is a human in the loop in our lines.
Rufus Grig
And I guess that speaks to the accountability, you know, that AI is a very helpful tool, but as humans and as organisations, we're absolutely accountable for what our organisations do. Similar question to you, Dan, what have you found in terms of both skilling, but also, you know, more broadly adoption of AI technology within Dunelm?
Dan Speed
Yeah, so we've actually all been through sort of a mandated AI compliance training. So I've had everyone in the business go through what to do and what not to do with AI. I say. There was a lot of that concern that Neil mentioned around compliance and concerns around what we're doing with our data and not wanting to expose the business into anything that we shouldn't be. Whilst there was a definite desire to be embracing AI and put it in, there was always that hesitation around which tools should we be using, what's safe to use, what isn't safe to use. So we got ahead of that really early and we had some support in making that sort of available to all of our colleagues and saying, This is the basics, so everyone's got a basic level of foundational knowledge. Not enough to go out and sort of go and create your own models or anything like that. It wasn't a technical training. It was more about just understanding the pitfalls of AI and how to use it properly. And then what tools are done and we're currently considering safe to use and we're allowing our colleagues to use. And then setting up a process for individuals to sort of submit their ideas for AI and for them to be evaluated from. the appropriately skilled people to say that that is a safe product to use. So we're at that level at the moment. We're trying to get the awareness out of how to safely use it, and we're developing up from there, really.
Rufus Grig
But you're creating an environment that encourages people to, firstly, understand what's available, but understand how you need to be able to put those guardrails, the responsibility, the data privacy, everything else alongside it.
Dan Speed
Yeah, we're very much a DevOps culture at Dunelm anyway. We've grown massively over our technology capabilities over the five years and there's no way that wouldn't have happened if we'd have been too sort of tightly controlling on what we can do. If we'd have gone in and said no absolutely you can't use that tool in, people would have found a way to use it anyway because they're all naturally curious. We actually we feel it's better to create a path for people to use with the appropriate guardrails as opposed to putting a fence and a gate up and saying you can't do it because people find the way over that gate if they wanted to anyway and the whole ethos of our platform team is to give people the guardrails to go out there and do whatever it is they need to do to make the business succeed, but for them to be able to do what it is they do in a safe environment. And that grows from our use of public cloud systems. And that's sort of naturally growing into AI. And it's how we keep the business safe, whilst allowing that experimentation that drives us forward and brings forward those technology advantages that we can then give to our customers.
Rufus Grig
Yeah, great. And then finally, Neil, I guess, how do you when you're working with retailers who might maybe be at the beginning of their their sort of ai journeys how do you encourage them to place those correct building blocks in place and what are the sort of the starting pointers well i mean just echoing some of the things the guys said around training and adoption i mean it is the number one reason why ai proof of concepts and pilot projects have failed is
Neil Holden
because there hasn't been that focus on people which is slightly ironic with ai implementations you know sometimes it's so obvious i think that um some businesses forget you have to train people but the training needs to focus on
Rufus Grig
literacy and confidence where ai is concerned not just technical skills in fact the the technical skills themselves are often taken care of by the models so there's less reliance on training people technically and more reliance on training people to be curious and inquisitive and want to give it a go and try and test and learn but that only comes through kind of literacy and confidence so If you're going to put an AI capability in front of a marketing team, for example, and say you can now speak to your customer data, unless you're going to encourage that curiosity about the types of questions that they can ask, then they'll say, OK, I might ask one question and then get the answer and leave it. Well, you're not going to get the full adoption then of that capability. They need to ask the right questions, smart questions and be curious in that way. But I think the biggest block around that is just fear of the unknown. This is... revolutionary stuff in some cases in other cases it's not but um i was speaking to someone earlier actually about this i i think i'm in a bit of a unique position where i'm old enough to be quite stuck in my ways and having seen you know the birth of mobile phones and all the other technological advancements of the past there is a bit of a fear factor even inherent in me and i run an ai business i should be well you know into this and behind it all there is a little bit of a change fear from myself going oh I'm not sure I like this Because it's just incredible the things that you can now do. But then hopefully I'm young enough to say, no, come on, let's get behind it and be curious. But training and adoption, massively underestimated, I think, and something needs to change in the industry if these investments are really going to take hold.
Neil Holden
Yeah. And I'm going to put you on the spot, Neil, and just ask one final question for you. You know, what do you see? What's the next big thing you see coming in the next sort of 12 to 18 months that's going to have a big impact on retail? AI associated with it unless you just have to say almost everything that's coming down the line has got AI associated with it these days is there one big thing if I had to pin you down one big thing
Rufus Grig
I don't think there is one big thing this is the part of the nature of AI there's going to be so many but I'm not that's not a cop-out I'm going to try and answer it I think the convergence of a number of things are kind of coming together I think the next 12 to 18 months will be about convergence things like voice commerce which actually isn't new it's been around for a while But AI-enabled voice commerce combined with AI-enabled search online, and that natural language processing kind of within that, I think is going to dramatically change the way that consumers shop and interact with businesses. Combine that then with home AI agents like Alexa and Google Home and so on, with voice commerce, that convergence of all of that, I think, you know, websites will websites be dead entirely in the future will that be within 12 to 18 months potentially will it all be done through voice and personal ai assistance at some point it will be whether it's next 12 to 18 months let's see that is a really bold prediction the death of the web from neil yes we'll look to see how that goes but i think yeah voice assistance and voice activation that'd be a really interesting thing to see one final question for you dan it.
Neil Holden
If you could change one thing about how retail approaches new technology development and AI in particular, what would it be?
Dan Speed
Firstly, just to expand on Neil's point, just to steal the question for a second. I saw a great demo actually of where you'd got an AI chatbot that had been asked to book a reservation for a restaurant. The restaurant was also using an AI chatbot to sort of service that. And in this demo, they detected the fact they were both AI chatbots and started communicating in their own. AI language that was meant to be more efficient. And as part of me was a little bit scared by that idea that they're now just developing a language to be able to communicate to make that more efficient. And that's going to be really interesting as part of that convergence, especially in sort of retail and hospitality, where you do see these AI systems blend together.
Neil Holden
Agent to agent is definitely a really hot topic at the moment. Yeah, definitely.
Dan Speed
It will be interesting to see that in the retail sector as well as I think our way of consuming stuff will change entirely and this will massively affect e-commerce because people won't go to browser website anymore they will ask their local home assistant to say I want to buy some green paint for my wall and it will go out and it will find some green paint and buy it from a retailer and that's something that's going to be really interesting to see how how retailers such as ourselves adopt that and embed that into our technologies and support it and how the world of sort of search and e-commerce completely changes as a result of that.
Neil Holden
Thanks, Dan. And then finally, Tim, I guess, you know, what's the one piece of advice that you would give to any leader starting their AI journey, but I guess particularly with your focus on retail?
Tim Walpole
I think it's a journey you need to start on. And I think the most important part is find a use case that you can deliver that you know is going to add some value. Don't try and pick off this massive thing that you'll never end up delivering. As Neil said, POCs just happen and nothing ever happens. So pick something small, something tangible, get some value out of it and then iterate on it and do it in a way that you're really thinking about it responsibly. responsibly, not only in terms of your ethical AI, responsible AI policy, but do it in a way where you look at the sustainability of it as well. So think about what you're doing, think about the carbon footprint of it, but do something that's going to add value.
Neil Holden
Thank you. Well, that has been a fascinating discussion, but I'm afraid that is where we must leave it. We are out of time. Thank you so much to all three of our guests. Thank you to Neil Holden from Ignite AI Partners.
Rufus Grig
Thank you very much. Thanks for having me, Rufus.
Neil Holden
Thank you. To Dan Speed from Dunelm.
Dan Speed
Thanks, Rufus. Thank you for having us.
Neil Holden
And Tim Walpole from the Retail Trust.
Tim Walpole
Thanks very much. Great to be here.
Neil Holden
And if you've been interested in anything that we've had to say, please do get in touch and tell us what you think. You can find out more about Curve and about AI in general by visiting Curve.com. And please do listen out for the next episode. You can subscribe, you can tell all your friends. But until then, thank you very much for listening. And until next time, goodbye.

Description

⭐ Special Episode! ⭐ Beyond the Checkout: How AI Is Redefining Retail from the Inside Out

Guests include:

Neil Holden (Ignite AI Partners): on why AI is the biggest shift retail has seen (yes, bigger than dot-com).
Dan Speed (Dunelm): on why your fancy AI is useless without rock-solid networking.
Tim Walpole (Retail Trust): on the link between AI, wellbeing, and your store's bottom line.

Expect straight talk on:

Real-world AI use cases in retail
Data ethics and the customer trust contract
AI-powered colleague support and wellbeing
Why infrastructure is the unsung hero of innovation
The not-so-distant future of voice commerce and AI assistants

If you want honest insight (with a side of wit) on what AI means for retail right now - have a listen!

Hosted on Ausha. See ausha.co/privacy-policy for more information.

Transcription

Rufus Grig
Hello and welcome to The Learning Curve, the podcast from Curve that delves into the latest developments in information technology and explores how organisations can put them to work for the good of their people, their customers, society and the planet. In our first series we had a look at all things AI and in particular generative AI and in this special edition of the podcast we're going to look very specifically at some of the ways that AI is reshaping retail. not just for customers and for customer experience, but also for a huge part of the workforce, the UK workforce that works in retail, and also for culture and sustainability in retail too. My name is Rufus Grigg, and today I'm joined by three very special guests, all from the retail sector themselves, who came together a couple of days before we're recording this podcast to lead a seminar on the use of AI in London. And they've very kindly agreed to come back and share some of their thoughts and insights with a wider audience on the learning curve. And I'm particularly grateful as I wasn't able to attend the event. So I'm actually hearing this for the first time. So let's meet the guests. First up is Neil Holden. Neil is a veteran, a very youthful veteran, I would say, of the world of retail technology, now offering the benefits of experience in a consulting capacity at Ignite AI Partners. Welcome, Neil.
Neil Holden
Hi, Rufus. Thanks for having me.
Rufus Grig
You've seen a lot of change in technology and retail in your time. Where would you class the current position of AI amongst those sort of major changes that you've witnessed?
Neil Holden
Blimey, yeah. There's definitely been quite a few. I mean, when I first started in retail, it was green screens and computers had not long been introduced into the workplace. And there's been many variations of changes since then. Dot com, online payments and so on. But I can't stress enough how I think... this latest revolution in technology is the most transformative. In fact, it's why I've given up a 27-year career in retail to go and get after it and hopefully help other businesses gain some advantages from it as well.
Rufus Grig
Well, it'll be great to hear some of your thoughts on that when we get stuck into it. Secondly, we have Dan Speed, Dan's networking technology lead at Dunelm, which is one of the largest homeware retailers in the UK and still growing at a really impressive rate. Great to have you with us, Dan.
Dan Speed
Hi, Rufus.
Rufus Grig
Dan, just for a moment. imagine there's someone listening to this podcast probably overseas who doesn't know Dunelm. Can you just give us a quick introduction to the business?
Dan Speed
Yeah, so Dunelm's a omnichannel retailer. We've got a store presence of 200 and growing and we also have a large online retail presence as well. We specialise predominantly in the homewares market so everything you can think of for your home you can get at Dunelm.
Rufus Grig
Brilliant, thank you Dan. Really good to have you with us and then our final guest brings a slightly different perspective. The perspective of the retail workforce. Tim Woolpole is Head of Product and Innovation at the Retail Trust. Tim, great to have you with us on the podcast.
Tim Walpole
Thanks for having me. Great to be here.
Rufus Grig
And for those who don't know the Retail Trust, could you just give us a brief overview of what you do and your mission?
Tim Walpole
I guess in summary, set up in 1832 to look after the well-being of people in the retail industry. And since then, the ethos hasn't changed. We're here to support. The well-being of colleagues in the retail industry. We support about 650,000 colleagues for about 220 retailers.
Rufus Grig
Brilliant. Well, thank you very much for joining. It'll be, I think, absolutely vital to hear the perspective of colleagues in this transformation as well as all the other stakeholders. If I can turn to you first, Neil, with your sort of broader perspective, you know, what sort of things should retailers be thinking about with AI in 2025?
Neil Holden
I think 2025 represents quite a shift, really. So... 2024, early 2024, was when I established the business Ignite AI Partners. And I think we were just entering into a market which was quite chaotic. In fact, I read a recent report which was explaining the dynamics of the AI industry from 2024 to 2025. And 2024 was deemed to be quite an explosion of different models and solutions and a very chaotic sort of landscape. I think 2025 represents. a bit of solidifying of the foundations of the AI industry. So models are maturing, actual results are being seen to be driven from the solutions. So I think from a retailer's perspective, I would have anticipated that last year I'd have been very confused. I think they probably still are quite confused, but I think it would be the uber confusion in 2024 about what's good, what's not. Coming from the retail industry, particularly working for businesses that are long established and have a heritage, De Devoe. jumping into the newest solutions that are just into the market they want to see that it's proven that it gives results that there's less risk around that as well in terms of that adoption ai is changing the game i think where that's concerned because a lot of retailers are going into things quite early and testing and trying to get results from them but even those that are a bit more risk averse. I think 2025 will be more. about, okay, we understand this a bit more now. We're probably working with partners who are helping us to understand that as well. Let's actually do some real implementations using AI.
Rufus Grig
Yeah. Okay. Interesting. So, I mean, I think the timing piece is interesting. I mean, if we think about the sorts of challenges that retailers are faced, you know, changing consumer buying habits, you know, movement away from high street, retailers' perennial challenges with margins and supply chain and everything else. Now, what sort of issues that retailers face are being addressed or are addressable by AI? Where are those applications?
Neil Holden
Yeah, well, there are many. It is vast. You know, we talked in the session about how AI could be seen as more of a utility like electricity than a bit of software, which is great in one sense, scary in another, a bit confusing in another. The real key, though, is to understand where those problems can be addressed using AI-related technologies. And if you just say it all can. that's not very helpful. You really want to pinpoint exactly where the biggest benefits are, where the biggest problems are that can be solved. Retail, as you just mentioned, you know, has faced many headwinds over the last five, 10, even longer years. You know, in more recent years, it's been about costs, you know, inflationary headwinds, consumer confidence, as you said, general cost of living crisis, and all of those things hit retail quite hard. Where AI can lend a practical hand to those things is around efficiencies. So things like intelligent automations, helping businesses to do things quicker and more effectively using AI-based technology. Also understanding your business and your customers in far greater depth than ever before and more quickly by harnessing the power of data with AI technology combined either within that or on the front end. So we see lots of our clients asking for things like. natural language interfaces into their data. I want to talk to my data. I want to understand things in natural language. And that solves a couple of problems. Firstly, there's less of hard data mining and data science required because a lot of the solutions can now do that. And so it democratizes that for people who aren't data scientists. I can still get great insights from data, but use AI technologies to achieve it. And secondly, to understand consumers at a much greater depth than ever before. Things like sentiment, you know. Call centers have long had analytics that say, this is how many calls you've had and what they were about and what you should maybe do next. But now AI brings things like sentiment analysis to the table far more quickly and effectively. So you can not only understand all of that, but you can also understand, how do my customers really feel about us as a business? How was their last interaction? Were they happy? Were they annoyed? Were they angry? Is there something that we need to do there to then follow up and recover that conversation? the vastness and the depth of capability that... that AI can provide for retailers is quite extensive.
Rufus Grig
Yeah, I think it's similar sorts of challenges that the businesses outside retail would be looking to AI to solve, I guess. I mean, one of the things we've talked about in other editions of the podcast is that the state of AI now, particularly things like generative AI and cloud-delivered AI, is that it is having a sort of democratizing effect in that where it was the preserve of the very largest of organizations that could afford the compute power. the the data scientists just the implementation costs it's now you know within the realms of much smaller organizations is that something that's happening in retail are you seeing a leveling or is there still a divide between those who can afford to use this sort of technology and those who can't there
Neil Holden
is still a divide for sure there's no getting away from that i mean the democratization of ai technology is nothing like i've ever seen before you know i can sit here in my dining room and get access to some very powerful technologies as an individual, not as an organization, and do some quite clever things with those. So if you then factor that into a smaller business, you can do similar things. But the divide is where doing proof of concepts or small projects using open source or free tools is one thing. But if you then want to scale that across your business to get real advantages from that over a longer time period, that's where the cost divide does come in. The leaders in retail, people like Curry's, who have really adopted AI technologies to a big degree and are seeing great successes from that, have an alliance with Microsoft and no idea what the budgets are, but I imagine they're very large. For a smaller business to do that to that degree would be cost prohibitive at the moment. However, the cost profiles have to change with the underlying technologies, and I think they will. I think it's inevitable. but at the moment costs can't continue to increase in line with the capabilities in a direct correlation. There will need to be a leveling off of that. And hopefully then that will give more advantage to small and medium-sized businesses to still get access to these great technologies, but not with exponential costs.
Rufus Grig
Yeah, okay. I mean, you talked a lot about data and the use of data and the ability to analyze that. We think about sort of the responsibility in AI and openness and explainability too. what's the sort of Contract with customers around the use of their data, your retail organization's use of their data. You know, how open are organizations being on that? And is it something that customers and consumers are going to be happy with?
Neil Holden
Yeah, that's a good question. So I think the industry is still figuring this out. Legislation and regulations are catching up, but not quick enough. Common sense prevails in most cases that businesses understand their consumers' concerns and requirements about how their data is used. I mean, that's not a new thing in GDPR and the right regulations are in place for that. What AI does is makes that 100 times more important because it puts a microscope on the potential misuse of data and how quickly that can turn into an actual issue because of AI is just ever so heightened. So businesses, I think, are trying to understand, well, we're doing all these things already in terms of data protection, information security, and ethical use of some of these tools. Is that correct? Is it not? How much do we need to tell people? Our advice to businesses is you can't be transparent enough. And I think if you are, and you're... behaving responsibly and ethically and and openly in terms of that and telling your customers where you're using ai where you're not how their data is stored allowing them to opt in in the correct ways then you shouldn't go wrong where that's concerned but we're often asked what new legislations and bits of regulation do we need now need to comply with and unfortunately i mean the eu ai act is the latest most up-to-date piece of legislation but it's I'd say it still has a load of gaps and it's not even enforceable until 2026, which I find a bit baffling.
Rufus Grig
Yeah, it's interesting, isn't it? The pace of technology development is outpacing the ability of legislators and regulators to remotely catch up and get ahead, which is a slightly troubling state of affairs. Thank you. If I could turn to Dan, I know that you were speaking when we were together at the event about how do we actually... place the groundwork to enable some of this technology to be able to be in use you know throughout the retail estate and we think about you know computing as being the important part of AI and the big number crunching and the GPUs and the data centers but actually if you can't connect if you can't actually access that compute or you don't get that connectivity in store none of this stuff is going to work so what's sort of the big problems that you were trying to to solve yeah so obviously I've got a slight bias in this area but for me networking is
Dan Speed
your absolute foundation to any of this from both sides so obviously you've got the networking that underpins the compute that generates and was working on the ai but equally if not more important is the networking in the area of where you're consuming it so for us in our store estate the foundation has to be there in you in your retail network you have to put the groundwork in and work with a good wi-fi provider to make sure you can put that in securely and to To the point around the experimentation with AI and within retail, there's a lot out there, there's a lot of tooling that retailers want to get their hand on and to start and see what that does. But there's obviously questions around new hardware coming into the estate and new devices that we want to be able to put in and we want to make that available and see what that can do for us. But obviously there's concerns around putting devices on the network that might have access to things they shouldn't so that security element it's foundational to being able to do that without having the platforms in sort of what we've done with Juniper most there's no way we'd be able to be in the position I if I think to where we were five years ago with the the hardware we had in the estate to be able to turn around to a one of the product needs and say yeah that's fine you can connect that device here here's a key and off you go we definitely wouldn't have been able to do that whereas now those conversations come and go a lot quicker let's just wind back a little bit you mentioned juniper um so you've
Rufus Grig
put this Juniper Mist Wi-Fi into the store estate. Tell us what's different and special about that and then perhaps it'd be really good to unpack the idea of some of the devices that you are wanting to connect in because you know for a lame person like me you know in-store Wi-Fi you know I tend to think of that as a thing that a customer connects to so that they're on their phone rather than something that's being used actively by the retailer to improve my customer experience?
Dan Speed
Yeah, so... The idea of being able to have those devices in the store, so we've looked at certain cameras, we've looked at devices that count customers coming in and out of the store, we could potentially be looking at systems that, as Neil mentioned, are gauging the mood of the customer, and we can be looking at services that support new technologies going in and around self-checkout. They're all devices that need something to be able to connect to the internet. most of the AI processing that's done in the cloud or it's done a service remote from the device because the amount of power that's needed to do that and to be able for that to happen you need to know that your device has got a constant feed to what it's trying to connect to so MIS has this idea of service level experience and that's really well tracked through the platform and it's using its AI and machine learning in the background to process what's going on with that and to be able to give us a very quick scorecard of what's going on with the estate.
Rufus Grig
So you're using AI to deliver the infrastructure over which you can then operate more AI. So eating ourselves here. That's really interesting. And I think I'm totally with you that that infrastructure, that connectivity layer and the reliability of it and your ability to just to be able to deploy things and innovate and experiment. And the answer that you can give back to the business more often is yes, which is, I guess, what you always want to be able to do, isn't it?
Dan Speed
Yeah, 100 percent. We I remember being in a world of networking where the most common response I gave to a question was no, or it's going to take a while. Now we're going from a point of yes in sectors where we've not had this level of investment and we've not had these excellent products we've got now. You've almost wanted to shy away from conversations because you know ultimately it's going to get to a no or it's a lot of work or you're going to be giving out disappointment. It's no longer a utility, it's a product. It's part of that journey of bringing something to life in our stores.
Rufus Grig
No, that's brilliant. Thanks very much, Dan. Tim, if I can turn to you for a minute. Dan mentioned what the Wi-Fi has enabled them to be able to do in terms of being able to equip colleagues in store with useful tooling. I think I'm probably guilty, as maybe some others, of not really grasping the vital importance of retail workers on all of our lives. You know, and it was the pandemic that really kind of brought into critical focus the fact that, you know, they're real frontline workers sort of exposed to the public all day, every day. Could you elaborate a little bit on how colleague well-being can sort of impact upon retail experience, on the performance of the retail business, the performance of the store, as well as on clearly, most importantly, their own lives and family situations too?
Tim Walpole
Yeah, I think it's a really interesting question that is probably not thought about. So if I take, I guess, an example of maybe someone that's a retail colleague and Someone has come into that store and either physically or verbally abused that person, which as we know happens on a daily basis to most colleagues. They go away and it's impacted their mental health, for instance, and they come to work to the next day and they really don't want to interact with customers because they've had the experience the day before, they're not feeling quite in the right state of mind. So they come to work not mentally prepared. And the effect on that is that the next customer comes in and they don't want to deal with them and they don't go out and say, hey, how can I help you today or something? It's just not I'm going to sit here. What does that do, of course, to the experience for the customer? And this negative experience, I walked into the shop, no one helped me. And how does that go? Well, that affects the store performance. It affects the financial performance. So you can already start seeing the experience from that customer, that mental health support or something just because of that one item of abuse. really impacts not only the colleague, but therefore the customer, and therefore the business as a whole. And at the Trust, we realise that's important. And we want to be able to say, how do we support you with anything that may have happened? So how do we understand, maybe through Juniper, maybe through understanding that a colleague has been abused, that we can reach out to them, we can support them with some proactive content to be able to say, here's something to support you, such that when you come to work the next day, you're going, well I've had the support I need and I can I can cope with that with that proactive support and that help.
Rufus Grig
So are you talking about colleagues use of things like the handheld devices that Dan talked about to be able to report you know instances of abuse or instances of challenge in in the workplace and are you able to use those same sort of devices to be able to you talked about pushing content to be able to perhaps support people in in that instance?
Tim Walpole
Yeah indeed so whether it's those handheld devices or some other signal that allows us to be able to identify that. a particular colleague has experienced abuse, but maybe also which other colleagues were near that may have also seen that abuse and therefore is potentially affected. And yes, we have a large amount of data in the retail trust. As I said, we've got about 650,000 colleagues on the system. We have a large amount of preventative content to be able to say, actually, we could then reach out to you automatically. So just by identifying that person, we can then reach out automatically saying by the way, do you know the trust is here to support you? Here's some proactive content, or potentially maybe you need some counselling, some in the moment support or something. So by having that signal from that retailer back into the trust, we can then reach out to them to be able to say, here's the content, here's how we can support you. I mean, I guess the other part of it is we can actually therefore measure the impact that this has, and we can show the impact that this has. At a couple of levels we look at absenteeism. We look at presenteeism, which is coming to work and not being in the right state of mind. And also Collie Churn and all these things have a major effect in the industry. But by bringing this content in to support them, we can show the impact that this has. back to that retailer, not only in financial performance, but also their own well-being. And the bigger impact is social and economic impact. We can show there is a, through happier colleagues, there is less absenteeism, happier customers, which actually leads to a positive impact to the GDP of the UK as well.
Rufus Grig
Yeah, clearly, absolutely vital. How do you deliver content? I mean, we tend to think of, if I picture a retail worker, they're in situ, They're on the shop floor. not sat at a computer terminal. They're also typically very busy responding all the time. How do you actually get content to people? How do you sort of schedule time for them to be able to have that support that they need in those environments?
Tim Walpole
Yeah, that's a really good question. And actually, we work as a B2B2C. So we actually send our content to the retailer, because the retailer, of course, has the ability to push that content out. But actually what we do is through the use of really clever technology, through the use of AI, we can understand all the colleagues in a retailer. So we can understand potentially that some colleagues in the north of England are struggling with abuse or financial support. And we can tailor campaigns back to the retailers. So we're just building out the latest version of our platform using agentic AI, one of the latest buzzwords. But what that does is it allows us to write campaigns on the behalf of a retailer, in their brand, in their tone of voice, because of course retailers are also time poor in being able to send some of this content out. So through the use of AI we write that campaign for them. So in this case here's a campaign to support colleagues in the north of England struggling with abuse or financial support. We will write that on their behalf and they will send it out to all their colleagues who they have the contact with. But we've identified already who the colleagues are, what they should do. And we then track that on the retailer's behalf saying, by the way, do you realize this is the impact that this campaign around this has? And this is a financial and economic impact that it's had just by doing that campaign.
Rufus Grig
So the AI is enabling you to, I guess, to turn these campaigns around more quickly, which means you can support more people, more retailers, more retail employees. Have you got any sort of idea of just how much the... You can quantify how much the move to using AI for this is extending your reach over and beyond using traditional tooling.
Tim Walpole
Indeed. So we used to write all campaigns by hand in the retail trust brand, of course. And then each retailer had to take that and turn it into their own brand. The challenge is, of course, if I was to do that for 220 retailers and I was to do three campaigns a month and I was to do it in their tone of voice, the amount of... staff that we would have to be able to do that is just non-justifiable so not only as you say is it allowing us to reach the correct colleagues with the right content in the right tone of voice every month or potentially twice a month targeted to what's important but it also says we can do in a way and as you said that is hits the right content and we've already shown through our trials that there's been a 50% uplift in people reading content through the retail trust, through targeting the right people, automatically in the right tone of voice. We're here to support more colleagues. We're here to support the better retail industry. We're improving the well-being. We're reducing absenteeism, presenteeism term, which is a good thing for people. Business in the wider economy, it's great for the retail industry as well.
Rufus Grig
Yeah. I mean, that's an amazing story of how the technology is helping, but also such a win-win in terms of the positive reinforcement cycle of being able to react and support people in those situations and the ultimately financial outcome as well as welfare outcome you get at the end. Was there any challenges you had, and I'm going to come to the others on this as well, around training or upskilling people to be able to adapt to using? I'm more thinking of your immediate colleagues in transitioning to using these AI tools.
Tim Walpole
Yeah, it's a really good question. And it's easy just to throw AI out and say it's going to do write stuff but it's really important to do it ethically and responsibly. making sure you're using the tooling correctly. And as part of that, our first insight is we need to be able to bring the business on board and we need to have the governance and policies in place. So the first thing that went in is that here is our ethical and responsible AI policy that says how we're going to use AI. Here is some training for everyone to understand what that policy is and what it means, the EU Act, et cetera, and as they change, being able to bring people on that journey. but also making sure that the content that we produce is actually suitable to be able to send out. So our policy says there's a human in the loop in every single conversation we send out. My normal example, and I talked about it the other day, we had a piece of content where it suggested that actually all the retail colleagues should just take CBD oil. It's probably not the right thing to do. The human in the loop picked it up. Yes, and it's really important to make sure that you use AI in a responsible way and you have a human in the loop. We need to make sure our brand is truthful and honest and trusted, and that is a human in the loop in our lines.
Rufus Grig
And I guess that speaks to the accountability, you know, that AI is a very helpful tool, but as humans and as organisations, we're absolutely accountable for what our organisations do. Similar question to you, Dan, what have you found in terms of both skilling, but also, you know, more broadly adoption of AI technology within Dunelm?
Dan Speed
Yeah, so we've actually all been through sort of a mandated AI compliance training. So I've had everyone in the business go through what to do and what not to do with AI. I say. There was a lot of that concern that Neil mentioned around compliance and concerns around what we're doing with our data and not wanting to expose the business into anything that we shouldn't be. Whilst there was a definite desire to be embracing AI and put it in, there was always that hesitation around which tools should we be using, what's safe to use, what isn't safe to use. So we got ahead of that really early and we had some support in making that sort of available to all of our colleagues and saying, This is the basics, so everyone's got a basic level of foundational knowledge. Not enough to go out and sort of go and create your own models or anything like that. It wasn't a technical training. It was more about just understanding the pitfalls of AI and how to use it properly. And then what tools are done and we're currently considering safe to use and we're allowing our colleagues to use. And then setting up a process for individuals to sort of submit their ideas for AI and for them to be evaluated from. the appropriately skilled people to say that that is a safe product to use. So we're at that level at the moment. We're trying to get the awareness out of how to safely use it, and we're developing up from there, really.
Rufus Grig
But you're creating an environment that encourages people to, firstly, understand what's available, but understand how you need to be able to put those guardrails, the responsibility, the data privacy, everything else alongside it.
Dan Speed
Yeah, we're very much a DevOps culture at Dunelm anyway. We've grown massively over our technology capabilities over the five years and there's no way that wouldn't have happened if we'd have been too sort of tightly controlling on what we can do. If we'd have gone in and said no absolutely you can't use that tool in, people would have found a way to use it anyway because they're all naturally curious. We actually we feel it's better to create a path for people to use with the appropriate guardrails as opposed to putting a fence and a gate up and saying you can't do it because people find the way over that gate if they wanted to anyway and the whole ethos of our platform team is to give people the guardrails to go out there and do whatever it is they need to do to make the business succeed, but for them to be able to do what it is they do in a safe environment. And that grows from our use of public cloud systems. And that's sort of naturally growing into AI. And it's how we keep the business safe, whilst allowing that experimentation that drives us forward and brings forward those technology advantages that we can then give to our customers.
Rufus Grig
Yeah, great. And then finally, Neil, I guess, how do you when you're working with retailers who might maybe be at the beginning of their their sort of ai journeys how do you encourage them to place those correct building blocks in place and what are the sort of the starting pointers well i mean just echoing some of the things the guys said around training and adoption i mean it is the number one reason why ai proof of concepts and pilot projects have failed is
Neil Holden
because there hasn't been that focus on people which is slightly ironic with ai implementations you know sometimes it's so obvious i think that um some businesses forget you have to train people but the training needs to focus on
Rufus Grig
literacy and confidence where ai is concerned not just technical skills in fact the the technical skills themselves are often taken care of by the models so there's less reliance on training people technically and more reliance on training people to be curious and inquisitive and want to give it a go and try and test and learn but that only comes through kind of literacy and confidence so If you're going to put an AI capability in front of a marketing team, for example, and say you can now speak to your customer data, unless you're going to encourage that curiosity about the types of questions that they can ask, then they'll say, OK, I might ask one question and then get the answer and leave it. Well, you're not going to get the full adoption then of that capability. They need to ask the right questions, smart questions and be curious in that way. But I think the biggest block around that is just fear of the unknown. This is... revolutionary stuff in some cases in other cases it's not but um i was speaking to someone earlier actually about this i i think i'm in a bit of a unique position where i'm old enough to be quite stuck in my ways and having seen you know the birth of mobile phones and all the other technological advancements of the past there is a bit of a fear factor even inherent in me and i run an ai business i should be well you know into this and behind it all there is a little bit of a change fear from myself going oh I'm not sure I like this Because it's just incredible the things that you can now do. But then hopefully I'm young enough to say, no, come on, let's get behind it and be curious. But training and adoption, massively underestimated, I think, and something needs to change in the industry if these investments are really going to take hold.
Neil Holden
Yeah. And I'm going to put you on the spot, Neil, and just ask one final question for you. You know, what do you see? What's the next big thing you see coming in the next sort of 12 to 18 months that's going to have a big impact on retail? AI associated with it unless you just have to say almost everything that's coming down the line has got AI associated with it these days is there one big thing if I had to pin you down one big thing
Rufus Grig
I don't think there is one big thing this is the part of the nature of AI there's going to be so many but I'm not that's not a cop-out I'm going to try and answer it I think the convergence of a number of things are kind of coming together I think the next 12 to 18 months will be about convergence things like voice commerce which actually isn't new it's been around for a while But AI-enabled voice commerce combined with AI-enabled search online, and that natural language processing kind of within that, I think is going to dramatically change the way that consumers shop and interact with businesses. Combine that then with home AI agents like Alexa and Google Home and so on, with voice commerce, that convergence of all of that, I think, you know, websites will websites be dead entirely in the future will that be within 12 to 18 months potentially will it all be done through voice and personal ai assistance at some point it will be whether it's next 12 to 18 months let's see that is a really bold prediction the death of the web from neil yes we'll look to see how that goes but i think yeah voice assistance and voice activation that'd be a really interesting thing to see one final question for you dan it.
Neil Holden
If you could change one thing about how retail approaches new technology development and AI in particular, what would it be?
Dan Speed
Firstly, just to expand on Neil's point, just to steal the question for a second. I saw a great demo actually of where you'd got an AI chatbot that had been asked to book a reservation for a restaurant. The restaurant was also using an AI chatbot to sort of service that. And in this demo, they detected the fact they were both AI chatbots and started communicating in their own. AI language that was meant to be more efficient. And as part of me was a little bit scared by that idea that they're now just developing a language to be able to communicate to make that more efficient. And that's going to be really interesting as part of that convergence, especially in sort of retail and hospitality, where you do see these AI systems blend together.
Neil Holden
Agent to agent is definitely a really hot topic at the moment. Yeah, definitely.
Dan Speed
It will be interesting to see that in the retail sector as well as I think our way of consuming stuff will change entirely and this will massively affect e-commerce because people won't go to browser website anymore they will ask their local home assistant to say I want to buy some green paint for my wall and it will go out and it will find some green paint and buy it from a retailer and that's something that's going to be really interesting to see how how retailers such as ourselves adopt that and embed that into our technologies and support it and how the world of sort of search and e-commerce completely changes as a result of that.
Neil Holden
Thanks, Dan. And then finally, Tim, I guess, you know, what's the one piece of advice that you would give to any leader starting their AI journey, but I guess particularly with your focus on retail?
Tim Walpole
I think it's a journey you need to start on. And I think the most important part is find a use case that you can deliver that you know is going to add some value. Don't try and pick off this massive thing that you'll never end up delivering. As Neil said, POCs just happen and nothing ever happens. So pick something small, something tangible, get some value out of it and then iterate on it and do it in a way that you're really thinking about it responsibly. responsibly, not only in terms of your ethical AI, responsible AI policy, but do it in a way where you look at the sustainability of it as well. So think about what you're doing, think about the carbon footprint of it, but do something that's going to add value.
Neil Holden
Thank you. Well, that has been a fascinating discussion, but I'm afraid that is where we must leave it. We are out of time. Thank you so much to all three of our guests. Thank you to Neil Holden from Ignite AI Partners.
Rufus Grig
Thank you very much. Thanks for having me, Rufus.
Neil Holden
Thank you. To Dan Speed from Dunelm.
Dan Speed
Thanks, Rufus. Thank you for having us.
Neil Holden
And Tim Walpole from the Retail Trust.
Tim Walpole
Thanks very much. Great to be here.
Neil Holden
And if you've been interested in anything that we've had to say, please do get in touch and tell us what you think. You can find out more about Curve and about AI in general by visiting Curve.com. And please do listen out for the next episode. You can subscribe, you can tell all your friends. But until then, thank you very much for listening. And until next time, goodbye.

Embed