Digitizing C&A: What It Actually Takes to Make AI Stick

Marcus: Okay, if I would say it's almost three parts, so let's get started. Good afternoon, everyone.

Sunny: Hey.

Marcus: That's another cool, very cool live talk for this sort of growing European retail community. I'm Marcus from Impact Analytics. We are software provider AI in merchandising space. And that was everything on advertising, mobile advertising. I want these kind of calls to have interactions with practitioners, topics which really matter, let me just admit one which really matter in reality. And today we have the pleasure to welcome Sunny, Sunny from cna. And we will talk about the importance of the human factor and what change means in reality. This is going to be a pretty real talk. Very poor on buzzwords. There will be a couple of areas or we might come a couple of areas. Just want to raise your attention where Sunny will not be able to touch because this is in terms of confidential information. But other than the rest, I'll leave this to Sunny. Obviously this is going to be a real exchange with real practitioner understanding. If you have questions during the call, please drop them into the chat. We will do our best to make reference or take sort of a spontaneous turn in the, in the call and then I would say let's kick it off. We have, I believe 20 people around Sunny. Pleasure to have you. For the very few people who do not know you, who are you? What is your legacy behind AI transformations?

Sunny: Thanks Marcus and thanks everyone for joining. So hopefully I hope I can make it practical for you and what Marcus said to me is make it as real as you can, the pitfalls, the learnings and you know, what's worked for me in my career. So a bit background around me is I started my career as an engineer. For the last decade or so I've been working in fashion, so retail fashion across numerous roles, product management, project management programs, and started a lot of the fashion experience at a company called Sainsbury's who you may or may not have heard of, like a multibillion pound retailer in the uk. They when I joined them, they had a clothing offline business at the time worth a billion and they didn't have any commerce presence. And my mandate there was to get them online and selling their two clothing, clothing business online. Sainsbury's then acquired a company called Argos, Argos, who you may or may not know being one of the biggest purple retailers in the uk. And the mandate I had there is how do I sell clothing on the Argos business and create a new business channel for them. So spent a long time doing that and then got A bit bored. Joined company called Levi's. So largest any company you probably have heard of. And there I was doing multiple things, leading a lot of the digital to market area. So spent probably five years looking at the whole planning transformation and that was everything from assortment creation, visualization, planning, allocation, replenishment and inventory management. So they were more of a build versus buy company. And we had a big investment into data science. So a lot of learnings from that experience. And for the last couple of years, two and a bit years I've been at cna and at CNA I'm helping drive the transformation of our planning portfolio. So that's where I'm working a lot with Impact analytics, who are one of our vendors that we partner with on some of the transformation. And a bit around why I'm at CNA is one, there's a real appetite in the business to really transform and sponsorship from the top to do it properly. And for me that combination is quite rare. So it's an exciting challenge. It's what really gets me up every day and just the opportunity to transform such a big business that's a bit around me and my background.

Marcus: Perfect. Let's talk people change and opportunity. The cna large, massive sort of brick and mortar footprint, multiple markets, sort of geographies, multiple categories. Where is the upside? Where is the upside via AI and what are the challenges?

Sunny: I think firstly, if you think about the scale of CNA as a business that we have over 1300 stores, over 17 countries and if you think about what that means from a customer, full print, 2 million customers walk into our stores every day and that's 2 million decisions on do I buy, do I walk out the store, how much do I buy? And the reality is, I think across our industry is a lot of the times when consumers walk into our stores or on our online businesses, sometimes they can't find their size, sometimes there's higher markdowns or you don't have a relevant assortment for them. And I think a big emphasis on for us is how do we make sure when consumers walk into our stores, whether it's in a hot climate store in Spain or it's a colder climate store, they have an assortment that's relevant for them. So I think for us the big opportunities, how do we make it stop being a coin flip? And consumers walk in and they don't care too much around the planning and the machine learning models behind what they see in the assortment. It's just they get the cohesive collection that they want and a big emphasis we're placing is not just the science on the planning transformation, but a bit on the art, the whole creative component. Making sure we're segmenting customer groups, we're clustering our stores in the right way as well. That's the biggest upside and that's really the focus that we're working through. The one part I would add as well, just on some of the operational realities, you asked me to make it practical, but where sometimes there's challenges and I think as most retailers and I guess as you all as senior leaders go through this journey, I think some of the things that really haven't that you need to watch out for as you go through this journey is one, I would say data quality. And it's a bit easier now with the agentic piece because it can handle a lot of the unstructured data. But particularly with your planning transformation, I think getting your data foundations right is really critical. So that's master data, it's hierarchy data and a lot of the data in retails like CNA that will set up prior to the 2000 and tens. The reality is a lot of our operational backbone is an Excel sheet. So part of what I would recommend is just take a stare down, make sure you have your bronze, silver, gold layers and you're really thinking about data and your semantic layers because I think that's important to get to good outcomes from your planning. And then just on the second piece I just wanted to add just on the opportunity, but also where you need to just where it can bite if you don't get it right is just making sure you're bringing the organization on the journey with you. I think behavioral change is probably one of the big things I learned in the last five to seven years through Levi's and also at CNA is there's one part like the technology, but making sure the change and the process comes along with you and you're building AI literacy or digital literacy across the organization. So hopefully that answered the question just a bit on where I see the opportunity. And there's tons of opportunity here at cna, but I really see a big opportunity for us to get a more relevant assortment to be a bit more accurate in how we plan, which is why we're going through some of this transformation.

Marcus: Yeah, wonderful. So maybe one more question in this, on this first chapter you spoke about opportunity and also a bit of pitfalls. Let's even go get more specific. What are your personal sunny lessons you have learned in the sense of prioritization? What actually what is important, how to drive the change in an organization and what not. Maybe also referring back to your previous context just to, for safety reasons. What have you learned did not work from a real sort of concrete measure perspective.

Sunny: Yeah, I've had a lot of these examples by the way, on AI transformations. I think the biggest pitfall is treating AI as that an AI transformation. I think these things are big business operational transformations and it can't just be treated as a tech only. And the one part that honestly hasn't worked for me in the past is going tech first, right. And saying, hey, I have an excellent model, my accuracy is, is, is amazing. But I'm trying to put that on top of an existing process. Why? Because you don't get the benefit. You have a great model and your map is excellent, but no one's going to use it. Right. So I think also just making sure that you're not treating this as a tech only transformation and you're bringing the business on the journey with you. And when we, I went through that experience and actually we had great models, great outputs, but teams were still behaving in the old way where they were looking at every single output, reviewing it, making the changes, and then you start to question what's the benefit there. I think the process and the workflow simplification really matters in that regard. And part of what we do and what hasn't worked in the past is not having enough emphasis on the user and the outcome that you're trying to drive. So I think it's important that as you go through your annual operating plans and you think about where do you want to invest, what are the problems that you're trying to solve in service to your company strategy and how are you going to measure success and then work backwards from that with the users that are involved and co create with them because oftentimes when you go away, you treat it as an IT project, you try and deliver it six months later in the hands of the business, you'll get resistance just to your point around bringing people on the journey and the users. A big part of what we try to do is co create with the teams. As I think about some of the transformations, it's not just a digital group working in silo with the technology teams. What we're doing is having change champions in the business, bringing them on the journey with us. Because as we co create these solutions with the users, a planner is more receptive to hearing from a planner around how things work. Rather than someone like me going in and say, hey, we can do your job in this way now. So I think building empathy for the people that you're solving for is also super important. And it's an area I could say I probably overlooked in the past.

Marcus: Interesting. There's also one anecdote or one sort of insight which applies to my background. But maybe at this point, any question, comments, remarks from you in the call here, Anyone wanted to share something or ask a question? No. Yeah, there was something. No, there's one example which actually fits to this. What Sunny just mentioned is having sort of internal. And this happened at my previous organization actually on the retailer side, when you encapsule sort of the digital team and sort of on purpose give this team more freedom and more speed and more decision rights. The managerial sort of excellence is to determine how far, how far away from the mother company or from the core company. Do you position this? This can be too close, then it's not fast enough. This can be too far, then this creates a culture on its own. Then the cool people are there and the boring ones remain. And this is not healthy. So this is not healthy. And this did not work, for example, it is not particularly well the human.

Sunny: Just to add to that, I think a big thing that really matters is executives are always going to have nice fancy ideas around like AI and agentic. But I think it's also important when you're working with the business teams is you really empathize and you explain the why, like how is it going to solve the problems and the pain points that they're experiencing? What is it going to mean for the consumer? What is the potential upside on like full price salary or some of those business metrics? And I think if you convince the people and you have a real problem to solve, it really helps. So I think that I just wanted to add that piece, but also just not making it a black box right where you're going away, you're building an AI model, showing it six months later. And one of the things I really learned, and this was more at Levi's, is iterative demos. Delivery really matters, right? Being inclusive and bringing people on the journey, but also trying to shift left where you can and trying to drive value quicker rather than just waiting 12 months until you can realize value. So just wanted to add those two points as well.

Marcus: Yeah, again, this is about sort of just exchanging in this network of really specific stuff, this co creation. Can you give this a bit more flavor on what this actually means in reality? How do you co create?

Sunny: So for us, co creation is one we're bringing the Teams on the journey with us, once we have a problem, we do problem discovery. We understand what the problems are, how they work in their workflow, but we're ideating with them. And one example jp who worked with me at Levi's is we launched a showroom at Levi's, right. Really successful product. But why it was successful is we had a cross functional group of merchandisers, planners, pd, all involved in the working group and we would sit down, we would co create, ideate together, we would test hypotheses, kill bad ideas fast. But it was bringing the people on the journey, they would contribute. They know their domain better than I do, they know their workflow and how they collaborate and what's not going to work. But the easy part, well, not the easy part. The good thing about the area that we work in on the planning transformation and the merchandising transformation, it's very easy to validate a good idea versus a bad idea because you're solving for internal users. Whereas if you think about like the millions of consumers for the E Commerce Channel, it's a bit more tricky. So I think when you're looking at planning transformations, co creation to me is just bringing people on the journey, validating the ideas with them, making sure they're signing off. We're prototyping and we're regularly getting their feedback and iterating on it.

Marcus: Great.

Sunny: Incentivizing them, making them feel like hey, they're part of the solution that we, that we built. Whether that's build or buy. I believe you can apply that approach to any.

Marcus: Yep, the, I guess everybody in here in this call has is either in the middle of a comparable transformation or went through this in some other roles. And the reason why I say this, probably everybody knows what I mean when I, when I say sort of the world is exciting and there's lots of energy at the kickoff day, at the kickoff workshop, maybe a little while afterwards, but then it starts to get exhausting. So after a little while, not everything continues to go right. And you as a leader, we as leaders need to keep the momentum. What are you doing? What are your recommendations, Sunny?

Sunny: Good question. And something we're still trying to grapple with, to be honest. But for me I think it comes down to three things. One is what's the end state that you're trying to build and how do you keep everyone going in the direction of the vision? And for us, when you think about what we're trying to build, it's not just a fancy forecasting tool or transforming our planning processes, it's how do we become more consumer centric? How do we stop the teams of having to create pivots and going through all these spreadsheets but actually having the data they need to drive better decisions. And I think the stat framing on where we want to be is just more consumer centric. Having the relevant assortment really keeps the teams grounded. I think the other part in momentum, particularly on transformations like ours, right, where you can be on 12 to 18 month transformations, I think also look at the opportunities, as I said, to shift left. Part of what we're looking at, yes, we have bigger initiatives in progress, but we're also looking at how can we get speed to value and how can we build confidence as we go through this journey. Because seeing is believing and once you start your kickoff, if people don't see anything for 18 months, they start to be like, okay, this has just been going on in the background. So a big part for us because the large transformations we have, the integration, the data alignment, the process, the change management, you can't really shortcut that work and it does take time. But there are opportunities for us to shift left and get products live. And we're looking at incremental delivery a lot. And one example we're launching with Impact analytics what they call their Monday smart module. And that's just an enabler where instead of looking at the what happened in your like your typical BI reports, we're now pivoting to more of why did it happen? How do I make a commercially positive action to drive a better decision? And that's a product we're able to get live in a matter of months where there's some of the bigger transformations might take years as an example. So I think that's important. And then thirdly, I would say just keep your momentum. I think the trust compounds free transparency and repetition. And by that what I mean is we are trying to be inclusive. We have regular demos, regular reviews, we communicate to the org, we have change agents and we show them what is, but also the failures. Right. And it builds that trust across the organization. So that's kind of the three things for me that we try to do. So since we kicked off like this transformation last year, we've been trying to maintain the momentum with those three points.

Marcus: Wonderful. Again, briefly checking in. Any questions, comments?

Sunny: Well, I would just say sunny, you know, you've almost mirrored everything as a partner that we would be recommending. And that speed to value and the high performing team getting you the energy to go through a hard journey is absolutely Essential and then maybe just add in the guidance specifically for leaders on their role to act as a shoulder to cry on and support, but also to be the unblocker. So their, their role as long as they understand what they have to unblock and there is trust in the hierarchy for people to come to them and ask for that unblock, I think it's gets you through a transformation so much quicker. Yeah, yeah. And for us, I think we've built a culture now where people are not afraid to speak up. Right. And particularly with the transformations, it's a lot of hey, how did you get to that number? Why should I trust it? And I think it's. We've created a culture where people are not afraid to voice their leaders or say, hey, we need stronger explainability on some of this. But I do think what you touched on is really important as well. Having the leader sponsorship really matters.

Marcus: Yep. Jp.

JP: Yeah, John, just want to add also. So in addition to the communication that Sanji mentioned on multiple level, targeted to audience to kind of the core people in the team, to the wider community as well, because even people that are not immediately impacted or included in the project, they should know what's going on. So it's not kind of work in isolation. It doesn't form that elite that you mentioned, Markus, but also the empathy, the human side, because it's not only about the trust in the results that Sunny mentioned. Why is that number like this and not like that? First of all, there is still an inherent fear in AI. AI will take my job. So it won't necessarily take the job because that's never entirely the purpose, but it will change the way things are done. And this is where it's super important to basically also kind of handhold a little bit the people saying you're part of the game to define how your role will look in the future. There is an advantage for you of not studying 150 different Excel sheets, but actually getting numbers presented that enable you to work more intelligently. But all this is also, it's a change not only from technology point of view, not only from a process point of view, but also from a mindset point of view. And this is I think one of the core fundamentals of a successful transition as well.

Marcus: Thanks for the contribution. Both the sort of. We are moving towards the end of the half hour. One question which is sort of your individual or very personal point of view for AI. What do you think is currently overrated and why?

Sunny: And again this is I guess a personal perspective and I will Just add this. I think what is overrated is that I see a lot and that the human should validate the outputs forever. And what I mean by that is, yes, AI is a smart thought. It's there to help you and enable the teams. But I think part of what I learned from my experience at Levi's is you do need to make behavioral change right where you want your teams to maybe pivot to more exception management as well. And if your forecast accuracy is great, why do I need someone to go in and review every single output? I think it's part of what I would say is there's cases where you need to trust as well, once you can build the confidence. Because otherwise, if you're just going through and you're reviewing and you're changing every recommendation, you haven't really transformed anything in my opinion. You probably just added another layer of complexity. So I think I read a lot around it's there as a smart start, and in a lot of cases it is. But I also think there's a moment, particularly with the agentic and some of the agents, where once your models get mature enough, you need to trust some of the models and be more exception driven. I think what is underrated, by the way, just in organizations is saying no and maintaining focus. I think particularly now with how technology is evolving, there's a temptation to just do everything right. I get every other day there's a vendor pitching ideas or someone coming up with new idea. I do think you just need to create focus and start to say no and focus on the key initiatives that really move the needle on some of the problems that you're trying to solve.

Marcus: Yeah, that's. I think that's. I guess everybody can resonate, sort of can relate to this saying no and staying focused, which is a challenge because there's a lot of noise in the market, of course, at the moment. So this is not an easy task. You're about to say something sunny.

Sunny: Yeah. I think the hard part is not the technology, by the way. Like, there's tons of tech out there. I think the hard part is really just focusing what moves the needle and then you assess what's the right vendor, the partner, the solution, and how do we solve this in the right way. Maybe it's AI, maybe it's some sort of other solution as well. But I think everyone now is drawing so much towards, like the solution first because of how, you know, the publicity around, hey, Claude can do my job and some of that stuff. But I do think it's really important that you do prioritize what's really going to move the needle and make you a bit more productive. Because the hard part isn't building the intelligence, it's getting embedded into a place where you can make smarter, commercially positive decisions as a business.

Marcus: The probably there's going to be the final question if we, if we. And you take one step back and sort of here also here there's some different sizes of companies represented in this call and uscna. You are sort of, definitely sort of have a relatively big scale to experience. And again taking one step back, is there some sort of minimum scale where you say as a market player, as a retailer, as a, as a fashion market participant, AI starts to be relevant and impactful if you are above a certain threshold. Does this exist or is this a misperception?

Sunny: In my opinion it's a misperception. I think if I was starting a company today as a one person company, AI would absolutely be in my thinking on how do I embed intelligence and optimize the way of working. I don't think you need comprehensive AI studios or data scientists to get started now. What I do think you do need to do though is just focus on what are you trying to achieve, what's your like P and L goals, what's your customer focus? And, and then let's think about how can I leverage some of the technology now to be a bit smarter in how I work. I spoke to a colleague from a bank last week and they have 87% of their Jira tickets all automated now. All the, all driven by prototypes and via users and it allows you to have that different, that discussion. So I wouldn't say you need to be a certain size to start using AI. I mean I used, I've been using it for years now but I do think you have to get started and if you don't, you'll be left behind. Yep.

Marcus: What a wonderful closing statement. Any, any, any final comments or questions from, from this group? If not, I hope you take something away. There is no sort of deeper agenda behind other than sharing and sharing a bit of insight and hopefully you take some, some, some thoughts away. So thanks for the interest. Thank you Sunny for actually sharing this from your perspective. Very personal and see you soon. Thanks everyone. Have a good business. Thanks.

JP: Good seeing you.

Marcus: Bye.