The digital thread has been "promised" for a decade now; a fully connected ecosystem where a single digital asset moves from concept through PLM, into 3D and AI-enhanced imagery, and out to the factory floor without ever breaking. That tension, between a decade of theoretical promise and the practical reality most digital product creation (DPC) teams are still living, was the starting point for our recent PI Spotlight.

Moderated by Joshua Young (Kalypso), the panel brought together Keala Stephan (Patagonia), Bhargava Ram Kummamuru (H&M), Karina Ochoa (High Life LLC), and Anne Lupas (ISA Sallmann AG) for an hour that moved from why one brand deliberately removed 3D from its own KPIs, to a pre-order campaign that sold out an underpant design before a single physical sample existed, to what it actually takes to get a supplier to trust a digital pattern over a physical one.

🎥 Watch the full discussion below & read on for the full write-up.

After a Decade of Promises, What's Actually Changed & What's Still Broken?

Ask a digital product creation team what's changed about the digital thread over the last ten years, and the answer isn't the technology; it's the excitement curve. For years, the industry treated the digital thread as an inevitable, beautiful, fully connected end state. That's given way to something more grounded. The industry realised it isn't in the 3D business, it's in the business of making money from physical products, and every digital investment has to be judged against that.

Where there's genuine optimism is in how far the use of digital assets has expanded. A decade ago, a 3D asset had one job. Today, Keala traces a much longer list of things a single asset now touches before a product ever ships.

What has really changed over the last ten years is that 3D models were primarily used for visualisation, but now those assets are informing product reviews, colour and print decisions, prototype reviews, sales conversations, digital showrooms, and now we have AI on top of that for enhanced imagery.
— Keala Stephan, Patagonia

Interoperability is the Industry's Holy Grail

What hasn't changed is that interoperability remains the industry's core unsolved problem. Systems that claim to connect do, technically, but the cost-to-value ratio is often so poor that the connection isn't worth making. Most organisations, Keala noted, have "islands of excellence connected by manual work": product data in PLM, digital assets somewhere else, visualisation in a third system, and factories still receiving information through email or spreadsheets.

So what gets called a digital thread is, in practice, often a series of disconnected handoffs, many of them dating back to pandemic-era 3D and digital asset management pilots that were adopted out of necessity but never fully connected across departments, geographies, or partners.

Just take any DAM solution you want and try to put in an MDL file or an OBJ file. You can't even view them. The colour formats are different, the render formats are different. You don't have this problem in automotive, with Rhino or Blender or Maya. That, for me, is always the holy grail.
— Bhargava Ram Kummamuru, H&M

The fix isn't one system trying to do everything. It's clarity about what each platform is actually responsible for (PLM manages product data, DAM manages assets, and so on), paired with a shared governance layer: a common definition of what a "complete" asset even is, followed consistently by internal teams and external partners alike.

Karina explained that nobody should be manually re-saving and renaming the same file across five platforms. Where automation can remove that friction, it should. Not because it's exciting, but because it gives time back to people who'd rather spend it on the work itself.

Purpose First, Technology Last

Purpose, then process, then people, then platform, and only then, technology. Reverse that order and you create complexity without value.

Technology should never be the starting point. It's tempting to jump onto something new, but the starting point should be understanding the desired business outcome and how the technology can enable that.
— Keala Stephan, Patagonia

Ram framed this through the analogy that you don't decide to eat and then go looking for a reason to be hungry. The methodology his team now uses formalises this as an "opportunity assessment" versus a "problem assessment". It's fine to start from a genuinely new technological capability, but the root cause of pursuing it should still trace back to a real process or people problem, not the other way around.

AI, in this framing, isn't a special case. It's simply the newest form of automation, and it earns its place the same way any tool does: by solving something that was actually broken.

Are Tool Limitations Constraining Creative Decisions?

DPC teams have, in the recent past, quietly pushed designers to learn 3D's full complexity rather than questioning whether that complexity was solving the right problem.

We stopped adding new skills due to tool limitations. Rather than utilising the true skill of a designer, which is creativity, we'd accepted that because the tools weren't good enough, we needed to put people through training programmes to learn every button, every workflow. Is that really simplifying design? That's not what they were hired for.
— Bhargava Ram Kummamuru, H&M

The fix isn't rejecting 3D. It's giving designers more entry points into it, some low fidelity, some full build, so the tool matches the moment rather than the other way around. That principle showed up differently at each company on the panel.

H&M gives designers access to sketching on iPads rather than requiring full-fidelity 3D builds for early concepting. At ISA Sallmann, the starting point is protected deliberately: designers begin from their own creativity and inspiration, with digital tools brought in only afterwards to support the idea they've already had.

My creative colleagues start with their own creativity at the beginning, getting inspired by pictures, movies, art, and so on. They create their own concept out of this, and the digital tools they're using are their little helpers to support them in fulfilling their task.
— Anne Lupas, ISA Sallmann AG

At High Life, Karina protects the same instinct differently: her team doesn't take away the parts of the process people already love, sketching, working in Illustrator, and instead built internal "AI champions" within departments, colleagues who understand the tools well enough to translate a pain point into a solution without every designer needing to become a digital expert themselves.

Who Owns the Digital Thread?

Embedding digital capability across every function that touches an asset, and figuring out who actually owns and funds that, turned out to be one of the more practically detailed parts of the conversation, prompted in part by an audience question about who administers these tools day to day and how teams justify the budget.

There's no single ownership model. At High Life, Karina's team earns trust before it asks for scale: testing internally first, moving into low-risk live projects, then building the training that lets a workflow leave the DPC team entirely.

We test internally first, really quickly, by modelling what a workflow will look like. Then we go fast into live projects, picking low-risk ones we're confident about. Through that we're able to iron out the workflow and develop a training, so it doesn't stay siloed with us. We want teams to feel empowered by the tools, not like they constantly have to go back to school.
— Karina Ochoa, High Life LLC

Budget follows the same staged logic. Karina's team owns ROI conversations directly but avoids asking for everything up front: goals set at the start of the year, issues researched across departments, and requests to leadership staged so the case keeps building rather than being made once.

At H&M, ownership is split by platform type rather than project stage. Ram's team owns everything specific to the design landscape end to end, including engineering, since his team sits inside the same unit as IT rather than working around it. Shared, company-wide tools like Adobe sit with a different part of the business entirely. What ties it together strategically is a KPI shift; Ram's team made a deliberate choice to stop treating 3D adoption itself as a target, judging initiatives on the outcome they unlock instead, since different domains such as print, product and pattern design don't all warrant the same blanket push.

At ISA Sallmann, ownership is distributed by specialty rather than centralised at all.

We started using 3D along the whole value chain, so everybody working with the assets has the same base: the same training, workflow, data structure, naming convention, file formats. These are developed together, so everybody knows what we're talking about. Beyond that shared base, everybody is specialised for their part of the process.
— Anne Lupas, ISA Sallmann AG

Whichever model a company lands on, Keala argued the real acceleration point is the same. Ownership matters less than whether digital capability is still seen as "an innovative initiative" belonging to one team, or has become business-owned work that stakeholders across functions understand and rely on.

I think it is critical to have experts in the business leading the journey, a DPC team or other types of experts. But when it becomes more of a business-owned body of work, that's where you really get some traction.
— Keala Stephan, Patagonia

Selling Before You Sew

Asked whether interoperable assets genuinely unlock selling before production begins, the panel's answer was an emphatic yes, backed by a live example. Anne described an underpant campaign at ISA Sallmann built entirely on 3D and AI-generated content. Three designs were put to a public vote, the campaign ran end-to-end on 3D with no physical samples, and customers could pre-order the winning design before it existed physically.

The campaign was a huge success and shows how valuable 3D and digital assets in general can be. You can promote something, or sell something, before you even have an actual physical product.
— Anne Lupas, ISA Sallmann AG

Keala backed the same idea around trust rather than technology. Physical samples have historically built stakeholder confidence because people trusted what they could touch, and a digital asset can build that same confidence. But only if it's accurate, connected, and repeatedly proven reliable through exposure over time.

Karina added a commercial angle: a well-built digital twin can resolve costing and pricing conversations far faster than waiting on physical samples, particularly when a pattern turns out to be more expensive to produce than the original estimate assumed.

The Supplier Handoff: Where the Thread Actually Breaks

The digital thread is only as strong as its weakest handoff, and that handoff usually happens at the factory door. Some vendors have gone all in on digital workflows; others haven't, and closing that gap takes real investment in shared SOPs, training, and treating suppliers as partners rather than recipients of an email.

As Ram shared on supplier resistance, it's not stubbornness, it's incentives. Suppliers run their own consumption calculations, and if they don't trust a brand's digital pattern, they'll quietly redo the maths manually, which shows up later as inconsistencies brands don't understand. The fix, he argued, isn't a mandate. It's giving suppliers a clear commercial reason to trust the digital handoff in the first place.

For Patagonia, that reasoning is explicitly tied to environmental impact, and Keala was candid that asking suppliers to trust a largely digital sampling process, even in service of that goal, is "no small ask." It takes a clear, consistent case for why, virtual quality standards suppliers can actually meet, and time.

Where AI Genuinely Changes Decisions & Where It Quietly Fails

Pressed on where AI has actually changed an outcome rather than just sped up a task, the panel was refreshingly unromantic about the technology. At Patagonia, some AI pilots simply haven't delivered; the value promised doesn't always materialise. Where it has worked is in enhanced imagery that lets the team have more robust creative conversations earlier, sometimes reducing the number of physical samples needed along the way.

ISA Sallmann used AI-generated imagery to solve a real gap in styles where physical prototypes didn't exist yet needed representation for selling. It worked well enough to become a genuine capability rather than an experiment.

AI isn't magic just because it's AI. It always has to solve a problem and make things better. That's the judge of any technology.
— Joshua Young, Kalypso

The Unsolved Problems

Asked what still keeps them up at night, the panel didn't reach for the technology at all. They reached for the seams between things: the points where one team's output becomes another's input, and where quality is hardest to hold steady.

For Ram, that seam sits between disciplines. The intersection points between print design, product design, pattern design, and the overall assortment or collection layer above them are still handed off between teams and tools rather than genuinely collaborated on, and closing that gap is where he sees the next real leap in fluidity from idea to manufactured output.

For Anne, the seam is quality itself. She can follow every technical detail correctly in 3D and AI output and still not always understand why the results vary, a problem the team hasn't cracked yet. Karina located the same pressure in time: delivering across every workflow while keeping quality high and communication clear, without people losing hours bouncing between platforms just to access an asset.

Keala's answer doubled as a description of the job itself. The challenge of changing process and ways of working while keeping the business running is permanent, and it should be. That's simply the complexity of the business. As she put it, using a phrase she said she reaches for often, the work is a matter of flying the plane while you're still building it.

Will Complex 3D Still Be Needed as AI Evolves?

Yes, and not because of nostalgia for the tool, but because of what it currently does that generative AI can't. 3D remains the environment where pattern, fit, material, grading, and construction actually get resolved. AI can generate a compelling image, but it can't yet produce a manufacturable pattern.

3D gives you the accuracy. You're working with your pattern, and it's exactly what you're seeing in 3D, super close to real life. I'm not yet trusting AI to select placement precisely. To get there, you'd have to teach it on your own company's assets.
— Karina Ochoa, High Life LLC

For Ram, the real question isn't generative AI versus 3D, it's what happens when agentic AI matures enough to execute the complex 3D tasks, like digital fitting and print placement, that currently require a person driving the software manually. In that framing, 3D as a discipline doesn't disappear; it may simply stop being something a person has to operate by hand.

For Karina, the ROI of 3D isn't only the assets it produces. Teams that work in 3D end up understanding their product more deeply, because the tool forces them to. Even where there's resistance to adopting it, people who stick with it tend to end up more confident in, and more in control of, the products they're making.

Let's not go back to being technology-driven. Look at the process, look at the purpose, look at what we actually need. We had the same debate between Illustrator and 3D once. Let's not turn 3D versus AI into another one.
— Bhargava Ram Kummamuru, H&M

The digital thread isn't a finished infrastructure project waiting to be switched on. It's an ongoing negotiation between purpose, people, and platforms that happens to run through some very stubborn file formats.

The organisations getting the most out of it aren't the ones with the most tools, but the ones who kept asking, at every stage, what problem they were actually trying to solve.

We'll be continuing this conversation, and much more, at our next Fashion Tech Shows in NYC, LA and Europe. Click the links below to find out more.