Empower the people to embrace automation, with Sarah Krasley of Shimmy

Hello and welcome to another episode of Fashion Made. I have the pleasure to be sat down with Sarah Krasley of Shimmy Technologies. Sarah, hello.

Hello.

How are you?

I'm good. It's nice to have you here in Brooklyn.

I know, it's a brilliant building, brilliant place. I couldn't be happier.

Good.

Sarah, you obviously run Shimmy Technologies, and a little bit of research told me that your goal is to help companies embrace automation. My question to you to start today off, and we'll dive straight into it, is how are apparel brands going to design differently when automation comes into play?

Well, you know, I don't think this dystopian future that we often see of the robots doing everything and millions and millions and millions of people out of work is going to happen right away. I hope not. Yeah, me too. I think there will be sort of an augmented age where there will be human labor as well as more automation coming into factories. And we're already seeing that. There's automatic fabric spreading, and 3D knitting, and automatic cutting. But with sewing, you start to get into interesting things when planning out how to produce some things. The big question on a lot of apparel brands' and manufacturers' minds is, will— how will those nodes within a supply chain change? Will there be sewing of different parts of an assembly in one country, and then it will be sent very close to where there's demand for the garment to be finished? That's anyone's guess. It's still early, but we're all trying to figure that out. But I think what will definitely happen is that there will be reshored production. And that will be a design challenge, right? This is going to change how long we budget in for the time to market. You know, sewbots will not be able to perform some of the activities that a human sewer can do. I don't think the systems that product development and design groups are using now is giving them enough feedback to allow them to do design for manufacturing and the multi-parameter decision-making that they're going to need to have to do in the design environment. I mean, I think too, there's so much valuable information on the factory floor, you know, being that Oh, this curve that you really love in this piece of assembly, which isn't really vital to the look and feel you're trying to achieve, really jams up sewing machines over and over and over again. Or it causes needles to break off and to be a risky situation for the sewers or for eventually the robots being idle. And I think all of those considerations, when you start to feed that back into the design process, and attach it to blocks and to legacy ledgers of information about these kind of pillars inside the apparel brands, you start to design in a different way, and it's quite interesting.

Okay, so automation will change the parameters of design.

It has to. You know, you can't tell a human sewer how to do something in the same way that you would tee up a job for for a robot to perform a task. It's different.

Yeah, is there anything to be learned from additive manufacturing there?

I think the exciting thing about additive is that you can start to really fluctuate materials. So what if you had a material that would change over time, that would be stiffer in some parts and more loose or rubbery in other parts? So while lattices are important for that, I think even just the ability to grow materials and to blend them, it will happen too.

Some amazing new materials out. I'm really looking forward to speaking to— we've got a few material scientists that we're going to be speaking to later in the series, which really shed some light on what's to come. Actually, with all of these materials because it's, as you said, it's not just structures but materials that can change in different environments.

Yes, and ones that can be brought back into manufacturing stream faster. You know, we're all talking about circularity and sustainability and material selection and the end of life condition of those materials is is so embedded in that.

Yeah. So that's probably a nice segue into how you got started with Shimmy Technologies. Was it just yourself?

Well, yes. And it was certainly not a very linear path. No. So I was leading an emerging technology group at Autodesk. And I was building software for mechanical engineers, industrial designers, factory owners. I just got interested in the apparel sector. So I started looking at the tools, CAD tools, simulation tools, data management tools, and realized that the things that were being built in the apparel sector were quite not as far along as what I was building at Autodesk and what I was used to using in more mechanical settings. So I learn by doing. And I decided to choose the hardest possible thing to learn with, which was a custom swimwear line for women.

Wow. That's throwing yourself in at the deep end.

Yes, pardon the pun. And we developed 5 blocks. We used Optitex. And we used some of the tension simulations to see— we'd sculpt a digital avatar that had the circumferences of our customers, and we'd assess fit that that way, and it really helped us see what that workflow was like and see what the industry might have, you know, up in its future as customers demand more personalization and customization.

Do you think that this ideology maybe of custom-fit clothing for everyone— and I say it's an ideology because it's certainly not here yet— and from my conversations with everyone, It seems to be a little bit further in the future than maybe some people are talking about. Can that ideology fit in with the fast fashion model? Can they coexist?

Well, I think apparel brands and manufacturers are going to have to borrow the ways fast fashion companies do things in order to produce custom clothing. There's no way that the antiquated systems that we use will ever get us to a point where we'll be able to be flexible agile enough to produce. Yeah.

Do you think it's agile enough as an industry yet?

Well, that remains to be seen. I think it's a very uncomfortable moment that we're in right now for a bunch of reasons. We have all this pressure for digitalization that's sort of driving us, and we're not driving it, because consumers are shopping differently. You know, they're going on social media and they're getting excited about a particular trend. The listeners can't see me, but I'm gesturing wildly with a pretend phone in my hand. And, you know, on the other side, now I'm doing kind of like a Frankenstein move of slow, rigid systems that are in place to fulfill those orders. And it's just, we're getting towards capacity now and it's not sustainable.

Yeah, I think the automotive industry where you've come from, because you started, you got this spark of an idea from a design of a car or something like that?

Well, actually, you know, it was from a T-Rex song.

Okay.

So one of my favorite songs is "Bang a Gong." Okay. And there's a line in that that says, she's built like a car. And I was sitting in this meeting looking at an industrial designer sculpting the hood of a car. I'd been listening to that record. And there were all these simulation technologies being used to assess, how will we have the nose of this car be aerodynamic enough to achieve propulsion with an electric battery? And I was thinking to myself, huh, I wonder if simulation technologies like that are used in garment engineering. And found some of the simulation technologies. But I think we're only at the very tip of what's possible in terms of predicting garment performance and fit before you develop a physical prototype.

What needs to change?

I think, you know, the approach that we're taking is that there needs to be greater interoperability between the applications that companies are using. There needs to be a data management system that's less linear. You know, PLM is just— you, you input things through these endless form fields, and then you shoot it over to the factory, and they get a PDF. And, you know, I mean, it's more nuanced than that, but, but in the majority of applications, it's very linear. And, you know, the way we are working is changing. We are going to be in global teams. We're going to be in multigenerational teams. You know, the roles and responsibilities of people working in our industry are going to change. And I think we'll outgrow these very linear data management systems quickly. So, the tools need to change, but we also need to begin flexing people's skill sets. Into new roles. And this is an area that we are doing a lot of work in and it's personally really exciting to me.

Yeah. That culture change is a massive component.

That's right.

Because the two cannot develop without one another.

Certainly.

The tools and the people need to grow together.

Yeah.

So how is Shimmy Technologies helping that?

Well, we've noticed, you know, just being outsiders to the industry, that there's quite a bit of manual work going on that just doesn't seem like it's in line with the brains that these people have. You know, the first time I saw a technical designer assess the fit of a garment in real life, in 3D, translate what was going on with gravity and cross pull into pattern corrections in a two-dimensional pattern, and then flip back into 3D, I was like, I don't know even many mechanical engineers at Tesla who'd be able to have a brain that's that powerful and able to do that. And then I noticed that these people are doing a ton of manual data entry into Excel sheets and things like that. And that's just ridiculous. And the time studies we've done have shown that that's a significant part of their weeks. And as, you know, we were talking about fast fashion and as more and more styles come into play and we begin to do these micro seasons, those SKUs are going up and these people's workloads is only increasing. So our technology automates the processes that mainly tech designers use to go from a sample coming in all the way through to that sample being approved. For production. And then on the other side, getting to, you know, your comments about workforce development and what's happening in terms of automation in manufacturing, we're noticing that brands want vendors to take more of the responsibility related to digital pattern making, to utilizing 3D in order to accommodate virtual fit, and reducing the number of physical prototypes. And in order to do that, they're going to need to develop a workforce who can handle that. And so many of those tools are in English, and they're not designed for people who are not digitally literate. So we've been developing a video game that teaches garment workers how to do digital pattern making in the language that they use in their country. So we're starting with Bangla, which is the language that, you know, the garment workers speak in Bangladesh. And we're in trials and pilots of that right now.

That sounds really cool. I want to play it now. Okay. Can we go and play it?

Do you speak Bangla?

I don't. I want to learn it just so I can play this game.

Yeah. So, We've had a long-term partnership with IBM. We've used Watson capabilities to automate measurement gathering. So for our swimwear line, we needed to get our customers' points of measure so that we were developing these avatars and delivering product that fits their unique bodies. And body scanning isn't pervasive yet, nor is it a great experience for people who've never seen their bodies in 3D before. So we opted to send our customers to dry cleaners or tailors who would act as the proxy. And we made this fun little animation that allowed them to do speech-to-text to capture measurements. And while that's just a sort of a half step in a process improvement, It— so many of our customers said this was so fun. This became this empowering experience, whereas before it was this very dry, not fun thing to hear what the circumference of your waist is. You know, so doing that gave us a lot of voice samples with heavy accents. So most of the tailors here in New York are not Americans. So— and Watson's— speech recognition was quite good. So we started to think, well, could that be a way that we could get over the barrier of many of these garment workers never using a computer before, or thinking that their life's work, or their life's worth is less than the price of a cheap laptop. So our interaction design mimics what they're using on their mobile phones. The video games that they're playing on their mobile phones and speech-to-text.

That's really cool. It's very interesting, making something more fun, especially with technology, seems to be a really important way to getting people to use it.

Well, also I think it's a pathway to more dignified work. I don't notice that the majority of work done in this industry seems dignified, nor do I see people having a whole lot of fun. And when you ask people, why did you get into this in the first place? They say, I want— I wanted to put people in clothing that fits them. And I really— I revel in the process of making something. And I want us to get back to that. And I think so much of that is just removing this work that, you know, nobody wants to do anyway. There are established artificial intelligence engines that can handle it. And let's make an interface that's light and fun and can travel with you. And that's why we founded this company. One of the fundamentals of Shimmy is that we are CAD and PLM agnostic. We will work with any vendor who is willing to work with us. And some are more amenable to that than others.

Of course.

And I think the ones who are amenable will be the ones that will still be in business.

You have to be, you know how annoyed everyone gets when Apple produces another charger or—

Right.

In electric cars, there are like 3 or 4 different types of charging plug.

Yes.

And electric cars only came out yesterday.

Exactly, exactly. Even, you know, I don't know if anyone listening to this will even remember this. I don't remember it firsthand, but I heard when fax machines started, you could only send a fax from like a Brother or a Mitsubishi fax machine to another Brother or Mitsubishi fax machine, which is outrageous.

I know.

But that's how it started. And the same is what we are experiencing. In architecture, you have a building information model. And the mechanical folks use it. The structural engineers use it. The conceptual artists use it. The construction people on the ground use it. It's all one single point of truth, and it interoperates with all kinds of things. And that's where our industry needs to go if we want to achieve the speed that we just have to.

Yeah. How do we do it?

Well, I think, you know, I'm encouraged by the work that the IEEE is doing to put some standards in place. I think that's really important. You know, you have companies like mine who are coming up. And while we might not be able to create the waves that a huge cruise ship does in a harbor, we can certainly talk about it, and our solutions can push the giants in the industry to, you know, provide more SDKs for developers and just allow for more ease in leveraging a model created in one platform in another. And, you know, I remember in LA as well seeing just steam coming out of people's ears talking about how much they're spending inside their companies, either with redundant work or with R&D initiatives to deal with this issue. And we need to hear more stories like that because that makes the people in my cohort, the technology providers, accountable.

I find the most interesting— panel discussions, and that's a bit of a misnomer, but the most interesting ones are when you put all of the solution providers on a panel together and just let them argue for an hour. That's always fun to watch, but actually provides some really good thought-provoking statements from them.

Yeah.

And I think the more that we bring people together in rooms and have it out and be honest and open with each other, the better for this industry.

I think that's true. And I think, you know, I'll do what I can as a newcomer to the industry and an outsider who has been through this exact digital transformation in buildings, in factories, in mechanical design. I'll do what I can to help too.

From your experience in all these other industries, Do you think there's anything that apparel can learn?

I hope the apparel industry will be able to see how different manufacturing segments have utilized simulation. You know, they don't just make a physical prototype of a car and see, how's this going to work out on the road? There's, you know, they do finite element analysis to make sure that a bracket is strong enough to do what the warranty says it's going to do. You know, there's all this work that we can do in that space. And I think by next year, we'll see every major apparel brand having a virtual fit initiative. But I think once that becomes more commonplace, we will be able to do more with simulation. But getting there is having more folks in positions like, We're seeing a lot of job descriptions for 3D operators and that type of work. So, you know, from a cultural perspective, that's really interesting. Are you pushing a tech designer to use these 3D tools? You know, it's all about trusting each other on a team, right? The creative designer's got to trust that the tech designer believes in what he or she laid out and the intent of the garment. And that tech designer needs to trust that the creative designer heard him or her. And that will be interesting to see the interplay of the tool in that scenario if neither of those people is the one actually doing the modeling work. That's pretty interesting.

It's going to require some really good rendering and visualization, especially of drape.

That's right. And we're doing some interesting work here in New York to build a pipeline for more 3D operators to come on those teams. And paying attention to not only the technical skills that they'll need to have, but also the soft skills of working in a multigenerational team. And there's a great article in B of F last this week that everyone is talking about right now about, surprise, millennials are interested in more than a paycheck. They're interested in career development and mentorship and—

They're just people like us.

Exactly, exactly.

You know, yeah, I find the term millennial a very strange thing, you know?

Yeah.

People just, they're just young people and a generation of people just the same as everyone else, but have grown up with a bit more technology.

Yeah, a lot more. And, you know, there was a woman from Levi's who spoke about this at a PI Apparel event a while back. And she said, try putting a PLM system in front of a Gen Z designer. You know, it's like, it looks like a 10-year-old website, and they're not going to use it. And I think I think that will be another force that we'll have to reckon with as well. Yes.

Yeah. Visualization of data certainly needs to improve on that side of things.

I know the Gen Zs that I employ do not even like using Outlook as our email service.

Really?

I do it because most of our customers use Outlook and I want to— a lot of it is scheduling. A lot of it is data flowing in and out of email. I do it to stay in the experience that they're having.

But yeah, well, isn't it funny the comparisons that you can draw between that and then 3D as well?

Exactly.

You know, people are using old systems just because their customers or their clients or their partners are using them as well.

Sure.

You know, the classic Excel argument.

Yeah.

Certainly plays a part in all of this.

Yes.

It's interesting you say about new job titles. 3D designers, modeling. When we started Py Apparel, I researched the industry till my fingers bled, and those job roles weren't really around then, and that was in 2013 when we launched this event. Now they're everywhere.

Yeah.

If you can use your crystal ball, what do you think the job titles might be in 5 years' time?

I think, um, what I don't want to see in job roles in the future is something being crafted around a tool. So, I don't want to see a job description that says, we're looking for a 3D operator. They must be proficient in Optitex, and they must do this, this, this, and this, and have it all be about the tool. You know, actually, Tabia and Nancy Johnson hosted a salon a few days ago talking about evolving job roles, and you should you should talk to her about this as well. But I think it's a mistake for employers to scope a job based on someone's acumen using a particular tool. It's much more than that. It's the way they think. It's the way that they problem solve. I think we're going to see teams similar to what goes on inside a design consultancy. Where you have a creative designer, a tech designer, and it's just fluid. And that team is working on things together in lieu of these sort of rigid roles that we have now. And then layered on top of that, the responsibilities of the vendor growing and filling in a lot more of the digital pattern making, 3D modeling, potentially even renderings and visualizations. Used for marketing and, you know, AR experiences for the consumers.

Brilliant answer. Who do you think is using AR well now? Do you know any examples of that?

I do. I really, really think Neha Singh is very cool. She runs a company called obsess. And she worked at Google and then worked at, I believe, at, uh, Condé Nast. And her, her issue is that e-commerce experiences are just these, like, white gridded boxes, and that's not the way people shop or think at all. So she's really excited to look at AR and VR as a way to have these evocative consumer experiences. I think that people want to see what they're gonna look like in a garment. And while you can do some things with photo wrapping and make that sort of a clunky AR experience, I think this is where the digital model is going to come into play, and this is going to be a major driver of for apparel brands to ramp up their efforts in digitalization. No question.

Do you think that might be one of the new jobs that comes on board?

Sure, yeah.

So it's more post-manufacturing.

Right. And a lot of that is not that difficult work, you know? It's a lot of manual data entry. And if we remove obstacles like language or location, I think we can redistribute labor in really interesting ways that hopefully doesn't leave people behind.

Yeah. Well, it sounds like Shimmy Technologies could be right at the forefront of that.

Here's hoping.

Yeah. Me too. Me too. Because it sounds like a great game and you've got to tell me when it comes out because I really want to play it.

Of course. Yeah, we're doing— the wonderful thing about working in New York is that We have every— people from every country in the world are here. So we're doing a beta this Sunday in Jackson Heights in Queens with a group of Bangladeshi women who've all worked in garment factories. So it'll be— you learn a lot in those circumstances.

Yeah, yeah, I'm sure.

So maybe we'll sneak you in.

Yeah. Please do. I don't know if I'll pass as a Bangladeshi woman, but I can— Maybe not. I'll grow my hair a bit more, maybe. So it was really an absolute pleasure to sit down and talk to you today.

Thanks for coming.

That's a pleasure.

It was great.

Yeah, brilliant.

All right.

Okay.

Thank you very much. Bye.