Intangible Versus Tangible Technology

Upcoming Event: DAM x DPC Unleashed

As digital product creation strategies mature, a familiar piece of enterprise infrastructure, with a limited audience, is taking on a heightened role – and reaching an expanded community. Digital asset management (DAM), once largely confined to downstream marketing workflows that trade in fixed-form assets, is now being asked to support a far broader, more complex set of use cases across product creation, especially where 3D is concerned.

This is already changing the way fashion thinks about an “asset,” shifting the emphasis away from static files whose usage on different channels needs to be optimised, and towards the idea of assets being assemblages of living objects whose utility is felt just as much in product creation as it is in marketing and communications.

Exploring this redefinition of digital assets, DAM processes and platforms, and the growing talent base that interacts with them all, is the focus of ‘DAM x DPC: Unleashed’, an upcoming webinar hosted by AVP on 12th May, bringing together industry voices — including our own Editor-in-Chief’s, as host — for a grounded look at where DAM fits within the next phase of digital transformation.

AI Adoption Is Far From A Straight Shot

From the “tale as old as time” department: technology diffusion and adoption is still not a straight shot, even in the era of AI. In fact, AI itself might yet prove to be on a completely different, and far less predictable uptake curve, than any technology step-change before it.

Yesterday, WalkMe (which is part of SAP) released the results of a survey, wherein they’d asked a global panel of more than 3,700 employees and executives, across industries, to describe their interactions with, and levels of confidence in, AI. 

This was the fifth time the company has run research with a similar objective, so, given that we’re now more than three years into the generative AI roll-out, we might expect to see progress. 

In practice, the stats almost universally demonstrate regression. In the headline findings, nearly 55% of workers had sidestepped their company-mandated drive to use AI in the past month, and had chosen to complete tasks manually instead. A third of workers simply hadn’t used AI at all.

These would be damning indicators in their own right, but perhaps the most telling findings concern the gulf between how employees view AI and how executives feel about it. According to the survey results, less than 10% of end users are willing to put mission-critical choices in the hands of AI, whereas more than 60% of executives are. A cynic would say that might just be because executives tend to be more insulated from the negative impacts of those choices, but The Interline doesn’t have any cynics on-staff to ask…

Close to 90% of executives are also convinced that the AI tools they’ve provided to workers are adequate, while around 20% of workers believe this to be true. And, tied to that, more than 80% of executives think AI is improving productivity, while workers say they’re wasting nearly a working day, every week, dealing with “digital frustrations,” which adds up to more than 50 lost days per year – a metric that’s up more than 40% year over year.

This is not, we have to point out, the first survey of enterprise AI adoption to report some measure of rot between the promise and the reality. The previous tentpole surveys have all been flawed in some methodological way, though. 

And while the incentives here are certainly a little wacky (a company that sells a digital adoption platform found jaw-dropping KPIs that could be improved by implementing a digital adoption platform, you say?!) there’s only so much that selection bias can do to keep colouring this kind of data before fashion, like other industries, needs to reckon with the idea that putting AI in front of people is looking like a fundamentally different proposition to even the most contentious prior pushes in enterprise technology.

The steel man counter-argument to this data, if The Interline was asked to make it, would be that a lot of those mandates to use AI (and a lot of those inflated executive expectations) are probably built on Copilot or Gemini as those tools are expressed within the productivity suites that the companies already use. 

If you’re a Microsoft house, you’re probably going to try and force people to use the AI assistant that comes bundled (or “bundled” in the form of premium pricing or additional token-burn) with Teams et al. If you’re a Google house, and Gemini is right there in the sidebar of your Docs, Sheets, Slides and so on, you’re incentivised to start shouldering your talent towards those tools. This is especially true since fashion, by and large, isn’t in the business of “token maxxing,” and brands certainly aren’t in a rush to allocate AI budgets to employees that can equal, or exceed, their salaries.

It doesn’t take much time browsing through Reddit, or talking to real fashion professionals, to know that people generally don’t like those models – at least not in the ways they automatically manifest themselves in enterprise productivity suites.

Had this survey refined its scope to look at Codex, or Claude Cowork, the results may have been different – since these appear to be the tools, and the models, that inspire individual devotion and are more likely to fall into the bracket of “shadow AI” – systems that workers bring in from home, and choose to work with over their employer-stipulated equivalents.

Neither does this survey separate AI usage into copilots and chatbots versus embedded AI, as integrated into new or existing applications. An “AI mandate” can just as easily mean an instruction to be prompt-first for everything, or encouragement to use new AI capabilities that have been architected into the platforms professionals are already using. And there’s a vast spectrum of different approaches and attitudes when it comes to making AI a meaningful or mature part of the current tech estate.

However you slice it, though, there’s an uncomfortably common picture emerging – one where even the most capable AI isn’t going to enjoy guaranteed uptake, because end users are simply choosing not to use it, either because they think it’s bad (or at least not the most expeditious route to a result) or because they are concerned about automating their own positions away.

This pushback, and some suggestions for how to sensitively approach it, will be covered in The Interline’s AI Report 2026, due later this spring.

And that same report will also aim to tackle another problem that’s equally visible this week: the persistent lack of hard metrics when it comes to measuring the positive impacts of AI. Analysis released to UBS subscribers midweek has made some headlines, since it suggests that AI will help apparel companies increase sales and profit margins.

But there’s a difference between suggesting and pointing to proof, and the same analysis leans into the notion that sales per employee have risen and that margins are potentially bouncing back – but it stops short of attributing any of this to AI rollouts in any quantifiable way. 

The write-up indicates that companies are being cagey about where and how they’re deploying AI, lest they give away a competitive edge, which is understandable.

Less sympathetic, The Interline believes, is the argument that AI is so big and so broad that it’s difficult – or impossible – to measure its return in isolated areas. This has never been, and isn’t likely to become, how companies measure return on investment. And a wide-format roll-out of a general purpose “force multiplier” is only valuable if the people whose force you want to multiply like using it.

Needless to say, data pertaining to the value fashion is getting from the money it spends on AI is thin on the ground. So this, too, is something The Interline wants to tackle in The AI Report 2026, and readers will have the chance to take part in a big industry survey (whose results will be reported in that publication) from next week.

But if the industry is struggling to understand where AI should live, and how valuable a house guest it’s set to become, this week provided a rosier outlook on a more tangible application of technology: onshore, on-demand production. The Interline’s readers will be familiar with Unspun, the technology company (which also began life by proving out its platform through running its own brand) that has been doggedly pursuing a vision for 3D weaving hardware and software that can slot into the supply chain, for the last decade.

It’s been nearly five years since Beth Esponette, Unspun’s CPO and one of its Co-Founders, wrote a piece for us titled “Designing A Way Out Of The Downward Spiral,” which argued that new methods of manufacturing were the industry’s best answer to the vicious cycle of making to predicted demands, in high volumes, and dealing with the bounce-back afterwards.

Yesterday, Unspun announced a new stage of its partnership with Walmart and other brands, with the retail giant signing a letter of support for Unspun’s plan to build out domestic production infrastructure in the USA. 

This might not be the mass-scale investment the alternative-production sector needs to pursue huge scale, but it does represent two other things: recognition that technology has a good chance of having a quantifiable impact in the real machinery of fashion, rather than just in the nebulous world of software; and an understanding that a world where international freight prices are governed by the caprices of a few caporegimes is not a world where the industry can continue to ignore the maturity and beyond-unit economics of making garments closer to home.