Welcome back to the Interline Podcast. Today’s topic is a big one. Where does AI sit in the design and development process, and how far does it cross over with 3D and digital product creation? We hear a lot of discussions focused on generative AI from a content point of view – generating images, video, text, even 3D models – but less about where it is and isn’t fitting into real product creation workflows and tangible digital transformation strategies.
To get into some of that, I’m joined for today’s show by Sean Lane, who is Director of Digital Innovation & AI at Vans, part of the VF Corporation. Sean’s been an influential figure in 3D and DPC for a long time, but as you’ll hear, his remit extends much further into tech enabled innovation today. And it’s a particularly pivotal time to have that kind of title.
So let’s get rolling.
The transcript below has been lightly edited.
Okay, Sean Lane, welcome to the Interline Podcast. It’s great to have you.
Hi Ben, thanks for having me on.
Not at all. Pleasure is all on this side of the table. Let’s start with some grounding. So yours is a pretty all encompassing title: Director of Digital Innovation and AI at Vans. Let’s talk a bit about what your remit actually is and what you’re actually working on day to day at Vans / VF Corporation.
You know, it’s funny sometimes – my title is misleading. A lot of people see me sitting within the e-comm area and I actually sit within design and innovation. And so when we’re talking about digital innovation, it really comes from that angle. Although I think when I look at innovation and I look at digital, it breaks down a lot of silos. So I end up working a lot of downstream activities as well. A lot of what I do is really looking at how we take designers and elevate the concepts and the designs that they’re creating, and how we visualise those in a different way – maybe storytell around those so that they can get their points across. In addition, I’m always looking a little bit forward and thinking about how we take something and make it maybe experiential or interactive or elevate something that can be used in sales or internal meetings or e-comm as well. So it kind of branches or bridges across quite a few different aspects of digital. I’ll do things around virtual try on, for example, and kind of play around with how we might do that. And those are types of things that could ultimately be used on e-comm.
However, we’re not doing a lot of that at the moment. The other part on the innovation side is always looking at how we advance in other technical areas like 3D printing, for example. And so we do a fair amount with that internally from a prototyping perspective. But then just looking at maybe externally, how could we take a 3D printed item to market?
And has AI always been a part of your job title or is that a recent addition?
No, it’s a recent addition. I’ve been in digital product creation, which is what I would say is more formally known as CAD development. And so that was where I started at Vans. And as AI has emerged onto the landscape rapidly, that has been a part that’s been added to my title and really shifted a lot of my thinking as well compared to what I’d been doing in the past with digital product creation and working more closely with vendors and factories.
‘Digital product creation’ is also the working title that we use for the whole 3D design simulation visualisation and then the ecosystem of tools, materials, experiences, processes and workflows and things that have grown up around that. I think it’s fair to say there is a lot of crossover and perhaps confusion, and circular conversations, happening around where AI intersects with DPC where it doesn’t. And I think that’s maybe one of the bigger questions we can get out of the way early.
As someone who’s worked on both, do you see DPC and AI as complimentary to one another or are they in tension? Because I’ve spent a fair bit of time analysing both myself. It does feel like there’s a really muddled space where the high skill, high accuracy and capability of 3D is meeting the lower ceiling and sheer accessibility of AI. But that’s just from my perspective and my ivory tower as somebody who gets to comment on these things rather than actually do them. From your angle, what’s really happening on the ground?
That whole topic is great. And it’s something that’s really hot right now as well. Having worked in digital product creation for the last decade and working with brands on helping them build roadmaps and set up infrastructure, the thing that was always a challenge is: I feel like I’m always selling the idea of digital. Not so much from an e-comm perspective, but more upstream when you’re looking at development and very traditional processes, like cut and sew, for example. Taking those and shifting them to digital is one aspect of it. And there’s a big foundation from a digital product creation perspective.
But the mindset is really the key thing; it’s getting people to think differently, to look at something differently and make decisions about it. That’s been a constant hurdle. It’s something that I’ve dealt with in four, five, six different brands that I’ve worked with. Some are better than others. Some move more quickly than others. And there are factors that I think contribute to that.
But what I love about AI coming into the landscape is that it’s starting to get everybody. It’s kind of democratizing the idea of creating digital assets. And so it’s getting everybody a little bit more involved in it. And from that perspective, people are getting more comfortable with it. They’re gaining some levels of expertise with it. So I think that’s helped the digital product landscape.
But I also think that digital product creation has been now kind of more contained. When I think about PLM systems, they try to do a lot. And digital product creation was trying to do more than I think it really was able to accomplish. And that accuracy and that quality – getting all of that right still requires that level of detail. But visualizing and getting a concept across and getting those ideas, applying a digital product creation process to that was heavy. It took a long time and it was more challenging. And so you couldn’t do as many. Now we can do more with AI. We can get the concept across and now shift that into needing quality and accuracy. We need fit. We need the other things that might be more conducive to what digital product creation provides.
You see AI then as almost the start of the funnel and then that’s your broadest canvas. And when you get down to needing to paint things in additional detail – as you said, fit, garment construction and so on – that’s where it becomes the realm of simulation, virtualisation and DPC.
Yeah, I think so. It enables the start to happen earlier from that visualisation perspective. I also think that it helps you kind of tighten in, get to a better level of quality with the product earlier, meaning that the visual is more accurate. And then you can take that and apply that. I think in the future – we’re not doing this today – apply that to a tech pack, for example, to give a better visual or even a digital asset that’s provided to the factory when they’re producing so they can see a nice version of that. I do think that AI is going to continue to expand rapidly. And so I see where even it’s going to chip away a little more at digital product creation. But I still think there’s elements of DPC that you can’t really replicate. And you may not be able to for several years with AI.
Let’s put that more bluntly then. What do we think 3D can do that AI can’t right now? Forget that couple year horizon. Today, what do we think 3D can do that AI can’t and vice versa? I think I’m starting to crystallise my own thoughts here on what is the preserve of 3D and DPC and what’s the preserve of AI. But I’m keen to get yours?
If you’re looking at it from a true fit, today, AI cannot really accomplish that, can’t give you a visual of what that’s going to look like on a person or a body. But if we started to train models around our patterns and our fits for our brand that we are aiming at and grading and all of that, I think then AI can transform even that part of the process where you’re honing in on getting accurate fit from a garment. But it’s not really there yet today. I know a lot of people working on it. And I haven’t seen anything yet that solves that problem using just AI.
No, I think that’s fair. For me, they seem like they’re running on parallel tracks, maybe divergent tracks in the long run. If the reason you use 3D/DPC is what you kind of hinted at and described earlier – getting to a visual, taking an idea and getting a representation of it that people can make one or more creative or commercial choices based on, whether those are in-house teams, wholesale partners, consumers, 3D is quite a heavy lifting way of doing that. It’s quite a circuitous way of doing that. And it seems sensible to say, is there a quicker way to do that, that is not such a burden on people?
Where it stops short is when you get to the level that you described where you’re starting to trade in and where product outcomes are more heavily determined by engineering principles and technical accuracy and things. It’s something I think AI is pushing into, but it’s not there yet.
Yeah, I would agree. We’re doing things like playing around with fabric weights and things like that and seeing how AI handles that. But the true simulation and accuracy isn’t quite there, although you can get some really nice visuals from it that can get pretty close. The other thing that I think is really interesting is that you can get so close in some cases that people think that’s a real product.
And then they start moving on to other decisions because they think, we’ve already moved past those decisions we needed to make, whether it’s colour or material or whatever. And now we need to start making decisions about what that’s going to look like in our line? And how are we going to sort this? And what accounts are we going to target? What consumers are we going to target? Things like that.
Yeah, there’s a whole philosophical question that I’ve been juggling about, representation and simulation. They’re not the same thing, but you can visually represent something enough times and with sufficient fidelity that in a lot of use cases it stands in for simulation. There’s a whole question to unpick about whether just being able to accurately and repeatedly make something that looks believable is the same as understanding and simulating it. But for different use cases, it’s almost an arbitrary distinction.
Yeah, you almost have to be really clear on what problem you’re trying to solve, what decision you’re making using the assets.
Yeah, that seems right. Now you touched on this, but I think the most interesting element of generative AI, at least in terms of how it’s being made available, is the sort of ubiquity and the lack of gatekeeping. I’ve been a card carrying member of the ‘3D is cool’ club for 15 years, but I’m not a 3D designer myself. And a lot of work has gone into making the DPC ecosystem easier, to get into institutional education, grassroots self-learning, and the software has become progressively easier to use.
But I don’t think the vision is for 3D tools directly, and 3D design development simulation to be for everyone. The idea is for the results to be for everybody and to benefit everybody, but for the tools to stay specialised. And that’s like the polar opposite of generative AI where workflows like sketch to image and the reverse, it’s all open to anyone and everyone. And we’ve all seen that in our personal lives as well. AI is the least gate kept technology revolution I think there is out there.
What does it look like from your point of view for fashion and footwear professionals to basically have completely unfettered, unlimited, universal access to AI, even in their personal lives, that they might be bringing some of those expectations into work? Because it’s not like 3D, where there is an on ramp and there is gatekeeping and people have to learn the ropes before they can get into it. AI is for everybody. And I think that probably changes the way that it rolls out and becomes adopted in big organisations when you have people bringing it in rather than you having to bring them up to speed on how to use it.
There’s a lot to unpack there. When I think about my own personal experiences and the things that I do, like photography, videography, and things like that I love, AI allows me to do more now that I couldn’t do in the past – which I absolutely love. But I’m not expertly trained in VFX. And so I am going to make mistakes. And a trained person might see that. An untrained eye might realise something’s not quite right, but not know what it is. I think that with AI. I’ve noticed so many things where we’ve had people, team members off the books creating stuff and then bringing it in and using it internally. There’s not a lot of risk there from an IP perspective, but more so for their own personal use on a presentation. And they don’t realise those nuances of awkwardness that are within that image until you really look at it closely. And they think, wow, I created this really cool thing. And then you look at it, and say, yeah, you did. But there’s a lot of inaccuracies in there that you may not have noticed. And so a trained eye can see those things.
I think I have a pretty well-trained eye, but I get on social media, and I’ll look at things on there, I’ll see an Instagram post and, my gosh, I realise it’s AI and it fooled me. It’s getting good. It’s getting really good. And solving some of those problems that the untrained eye can’t see. And then I think about photography. I know lighting, for example. I know things like that. And so I know what parameters to prompt, I can go in and make sure I want this type of lighting, I want this focal point. I want this and that. And I can put that in a prompt, and it recognises it. Whereas somebody who doesn’t have that training would not be able to do that. It’s really hard to police. Sometimes not necessary to police. And then other times it is. And I think that’s where you have to make sure people are aware of the impact of doing this with something like a forward-looking product that we don’t want shared broadly or used to train a model, for example.
And I think the key bit there is sort of raising the floor of what people can do individually, but still making sure that the output of that goes through people with not just a discerning eye or taste, but specialists in particular domains, who are the ones able to distinguish between something that passes muster visually and then something that is complete and accurate enough to carry forward.
And I’ve come across this in my personal and professional life as well. I’ve spent 16 years as a writer. I do have some formal training in that area. I think it’s super cool that the average person can be a better writer, thanks to this. I think it’s interesting for people who maybe wanted to write and couldn’t for whatever reason, or found that the initial barrier was too high. But I also have the strange experience of being able to spot most of it within my domain. Can I spot it within everybody else’s domain? No. So visuals are not my area. I’m not a trained artist in 3D, 2D or otherwise. And I think this is where people fall into thinking about this slightly from the wrong way, which is, eventually, AI will trip me up, and I will read something written by AI, and I won’t know. It does seem like that’s the inevitable trajectory.
You know, my daughter is a writer and, man, that’s an area that is and is going to be heavily impacted. When we adopt technology, we don’t go backwards, we go forwards. And, the funny thing is, we’ve been using AI for a long time. A lot of people don’t realise that, you know, it’s machine learning or some form of AI that they’ve been interacting with, talking to and asking questions of. And now it’s just more prolific. But there are industries that are going to get impacted. And, how do we evolve with that? The other thing, too, is I think there will be a community of people that will want to have craft and want to know that something was developed by hand, by mind, by, you know, a human. I think about in Japan where they have factories that old school, you know, build denim and it’s still something that’s reveled around the world. And so I think that there’s still a place for it. But it’s going to minimise or impact industries that have that creative spark.
Yeah. We very much trade on and rely on the idea that long-form, heavy human touch content is still valuable. But there’s a whole media ecosystem that is built around the exact opposite idea, which is that cheap, fast and light touch is what people value. And I can’t argue with that. And there are some parallels between what’s happening in media there and what’s happening in fashion for sure.
As someone who’s been through and led kind of multiple rounds of enterprise digital transformation, does what we’ve just talked about there – you know, this idea that AI is a wave of transformation that’s coming for every industry – make this moment feel fundamentally different to you, or is it more of the same? Because I’ve struggled with this, reconciling the way people talk about AI as a platform shift versus the everyday reality of choosing and implementing discrete, narrow applications, integrating systems, working towards adoption through onboarding and training and so on. It’s like, there’s this big wide open possibility horizon and then there’s the everyday mundane reality of software.
When you break it down, this is probably no different than any other large transformation in the idea that it still requires people to adopt it, to make the change, and to interact with it. I do think that it is going to move at a much more rapid pace. And the brands that are not engaged with it today, I feel like they’re missing out. It’s not like digital product creation. It’s different. It’s not the same as, hey, we have to build this whole foundational infrastructure to be able to actually leverage this. You really can leverage this at scale in different areas today. And if you’re a brand and you’re not doing that, you’re going to fall behind.
I think that’s something that at VF I appreciate: they’re open to looking into and in some cases adopting different parts of this technology. And then there’s a lot of play that’s happening too that’s enabled. And I think that’s where if you’re a brand, I would enable play, enable your employees to have access to things that they can play with, but put guardrails and make sure that they’re aware that there is a potential impact to the brand through it. I was thinking about Patagonia as an example, it’s a brand that is environmentally conscious. And where does AI sit within that? Like what’s the impact? And even at VF, we’re environmentally conscious and have things that we want to accomplish in that space and where does AI sit within a brand to think about sustainability in a different way?
There’s a lot of different considerations to bring into this. And it’s strange when you mention the play side of stuff, they have the opportunity to bring in applications and experiment with them. There’s no shortage of applications.
If there’s one thing we found with a couple of years of AI reports under our belts it’s that people are very happy to make tools for all manner of different use cases in this, under this umbrella. Some of them are advancing further than others. Some areas are more mature than others, but if you want to try and implement AI in one or more areas of your workflow, somebody out there will sell you a platform to try and do it. So I think the weird thing is not to think about whether there’s a tool and to start with that tool and try and find a problem for it, but rather to say, what is the use case? What is the problem? What are we as a brand working to accomplish? What are the values and what are the kind of core attributes that define us? Can we deploy AI in service of those?
Yeah, this landscape is evolving so fast that I’ll have somebody come up to me and ask me specifically about a new thing. If you heard of this and sometimes I’ve heard about it, I don’t know a lot about it or I haven’t heard about it. Or yeah, I’m using it, you know, it just depends on where they’re at in the journey and what they’ve been listening to or looking at. But I almost feel like you have to have that inventory per se of problems you’re trying to solve. Because the other thing that I’ve found is as you get in and work with some of these tools, you might find that it can solve multiple problems. And so how do you say, I’m going to go after this problem, but look, it actually also can accomplish this. And that’s what we’ve been finding a little bit with some of the play that we’ve been doing to test out: how far can it go, and what does it do well, and what doesn’t it do well? And every single one of these models and subscriptions and so forth handle things differently. And whether it takes the prompt really well, or it takes an image really well, it learns from what you put in really well, all of those things impact the outputs. So it’s been a fun process, but it’s also very frustrating at times, because I’m like: I want to accomplish this and solve this problem, and this is not doing it for me, and I’m getting frustrated by continually trying.
And I think what’s maybe under discussed, is, just how wide that possibility space is and how novel that is because you mentioned PLM earlier, we talked about DPC, you know, even though the footprints of those things have been stretched in different directions, depending on who’s creating them and who’s implementing them, we’ve arrived at a point where it’s pretty well circumscribed what those things are supposed to do and where they begin and where they end. Having the chance to bring something into an organization where its shape is very much not fixed, feels like it would be jointly empowering and annoying depending on where you stand.
Yeah, definitely. I find a lot of it is empowering because it does enable things. I mean, I’ve been able to accomplish things that I wouldn’t have been able to otherwise, and do them faster. But I also have to accept, especially when I was first starting on this journey, the tools weren’t quite as far along. And so you would put something in and you would get something pretty wild back – a lot of hallucinations, a lot of inaccuracies. You might be somewhere in the process where you want it to look a certain way and you can’t get that specifically. And so I’ve brought in tools which enable a little bit more editing and capability. But then I’ve also seen the landscape evolve quickly. The tool that I brought in where I could do the editing and capability isn’t as strong as the tool that just came out and I can do it better now using that. And that’s hard to stay on top of. I think that’s the hard part.
Yeah, it definitely seems like there’s a big churn of new capabilities, which makes it hard to target something. It makes it hard to target any kind of individual specific initiative and then frame it based on the success afterwards. The other thing to think about is every technology rollout comes with cultural considerations and change management considerations, but I think AI is particularly acute in that area because people’s perspectives can be pretty polarized. And I think people are going in with preconceived ideas at a broad cultural level. Everybody has an opinion about AI, which is certainly not the case about your average person on the street. As interesting as PLM and DPC are to you and me, the average person on the street doesn’t have an opinion about those.
What does it look like from your point of view to get designers, engineers, merchandisers, marketing teams, etc. to feel like they understand the vision for transformation here and they can help influence its direction instead of just feeling like, look, AI is coming, it’s washing over me, it’s an inevitability and I feel some type of way about it?
I think the breadth of what is out there is so much more significant than we’ve ever had that it actually touches all these areas. When you looked at digital product creation, you had a handful of tools. You had a handful of options. And it was all about, hey, let’s do a bake-off to test and figure out what the best option is for what we’re trying to accomplish or what problem we’re trying to solve. And I think with AI, there might be five of those tools that work really great at solving a specific problem. And they do that really well for design. But then there’s another five tools that do that really well for, I don’t know, consumer insights. And then there’s another five tools that do that really well for planning. And then there’s 25 tools that do that really well for e-comm and for merchandising and for sales. And when you start putting all of that together, it’s the transformation piece. It’s going to happen, but how do you gain some level of control over that as an organization, especially a multi-brand organization? And so you have to have an element of governance. You have to have an element of people focused on transformation. You have to have people within, that are doing the work, actually filtering that work up to say, these are things that we can accomplish by doing this.
There’s so much exploration that’s happening, that there’s sometimes a disconnect in the speed of exploration to the actual implementation and transformation from above. And so it’s keeping those in sync. It’s probably a little bit more of a challenge than it has been in the past. Because in the past, you might have adopted a tool or two tools or three tools and said, OK, these are the core. It’s not like that right now.
Do you find that you have to be reactive to the way people feel about AI when it comes to those things?
Yeah, I mean, you’re still building digital mindsets. I think that’s a big part of it. I come in with a full on digital mindset and I’ll say, we could do this and this and this. And there’s still a response of, we can still do it this way. And so you’re still convincing people to some extent, or showing them or leading them to what the possibilities are.
The other part is fear. I think everybody has a little bit of fear of what the ultimate impact of AI is going to be, not just on their jobs, but their lives and their kids and future generations. As a parent, I think about that a lot. You know, what does life look like with AI as such a big player? Because it’s going to be, and it’s just going to shift the way that we do things and the way that we think.
I was thinking about this today. You have people who used to be expertise, used to be a lot of expertise, used to be here in the US around pattern and pattern making. And a lot of that’s been shipped out of the US. And once that’s gone, getting that back is not an easy task. And maybe sometimes it doesn’t even make sense. But if you think about design and creativity, once that’s gone, it’s really hard to get back. And so if you’re an organization who’s not being thoughtful about how you’re implementing AI, the impact might be detrimental long term. I think that’s super important to factor into all of your transformations – when it comes to AI, especially.
I would agree with that. We are an industry that has watched a lot of expertise leave, or at least be relocated. And when we’re looking at a transformation that has the potential to impact as many areas as we’ve talked about, the threat surface of that is much wider. And, as a parent myself, I also think about these things. I come at it from the point of view of which careers remain viable, not just from a pure automation perspective. But even if the automation doesn’t work out, even if we predict forward and it actually turns out to not be the case that AI could replace a particular area, did we go sufficiently far down the road that the pathways for acquiring those skills and the ecosystems and the opportunities for building those skills went away? And all of a sudden you end up with disciplines and domains that are just not as easy to access, not as available as they were before, and that is something to be concerned about.
Yeah, what skills are necessary in the future? It’s kind of a bigger question.
Absolutely. And probably beyond the remit of a short podcast, I think. So one we can unpick in the future.
Now, as an industry we talk about digital lot. We are still trading in excellence in physical products. For all the backend work, product outcomes are what determine success. Those outcomes, you can measure those in terms of fit, style, performance, values. There’s a lot of different ways the market will judge the end result depending on brand positioning, people’s priorities and so on. But any investment in technology or process innovation is eventually going to be graded based on how effectively it contributes to the outcomes that matter to brands and their consumers. Someone in your role with broad oversight of a lot of technology and digital and innovation initiatives. What are the things that you see, digitally speaking, contributing the most to those kinds of product outcomes today and how do you think that mix might change in the next couple of years?
That’s a great call out. If the product that you put into the market for your consumers is not hitting the mark or it’s not delivering on what the expectations are of that consumer, I think you’ve failed. So, for me, I look at every aspect of what I do. Does this provide value to the organization? And I kind of measure value in a couple of different ways. One is does it increase revenue? Does it improve margin? Decrease cost. And if it does one of those three things, then it’s worth spending some time on. There’s also maybe less tangible value propositions as well. And so those have to get factored in. Is it creating a better work environment? Is it creating excitement and energy inside the brand? Those are other things to consider as well. But ultimately, that product at the end has to be the right product.
There’s two things that I think it does. One of the things that we have to consider is like, sure, it’s going to solve a problem, but what other problems is it potentially creating? And I don’t think we do that enough. I think we can go and solve a problem, without realising I created three other problems, maybe downstream or adjacent to me, that now have to be solved. And so you really have to think about that part of the process as well. I do think that AI will enable us to be more focused on the things that create more value, because some of those redundant tasks we can put agents in place to take care of, or we can have processes that deploy AI in a different way to maybe give us recommended color ups and things like that, whatever it might be. And maybe the other thing, too, is that it can call out where you have discrepancies that you may not have noticed. It can do that analysis much faster than a person could. So those are value add, in my opinion, and that enables a certain role to maybe focus a lot more on what is important and adds value long term.
Amazing. Final question from me. We will be doing our AI Report again next year. It feels a little far enough away. And AI moves fast enough that I don’t want to commit to anything that’s going to be in it right this minute. We are, though, currently hard at work on the DPC Report for this year, which will be our fourth of those due out in December, just ahead of the holidays.
…which are all great reports.
Thank you. I appreciate that very much. Without giving too much away about anything that we’re covering at the moment, is there anything that you would like to see us put in there? Is there anything when it comes to 3D digital product creation that either isn’t talked enough about at the moment or is talked about but not at the right level?
I struggle a little bit with this because I have been a proponent of it. I see the value in digital product creation and the realities of it are not talked about enough. You know, I think there’s success stories, but when you look across quite a few brands, there’s also a lot of stories that have stopped and started many times. And I think that’s kind of an interesting discussion point to come at it from an angle of, I don’t know if failure is the right word, but less than success based on how you defined it. I think that’s interesting. What are the metrics, maybe, that define success? That could be a good one. And then this whole landscape around, how do AI and 3D kind of play together. I think that’s really important.
I think there’s two things we can pick up there and hopefully we’re able to do them both justice. For now, Sean, it’s been wonderful having you on the show. Thank you so much for joining.
Thank you, Ben. Thanks for having me on. I’m passionate about these topics. And I’m happy to talk more about any of them.
Absolutely well, we hope to have you back on at some point in the future then.
Great, thank you so much.
