Our regular analysis selects one or more news stories from fashion technology, and presents The Interline‘s take on why they matter to our global brand and retail audience – as well as what they might mean for the longer-term future of fashion. As always, this analysis is also delivered to Interline Insiders by email and signing up continues to be the best way to get a fresh look at the fashion technology news, completely free, in your inbox.

AI chatbots and customer consultations: retail’s different approaches to product discovery through technology

In a rush towards something of an acronym soup of hot technology topics, French luxury group Kering announced this week that it would be harnessing the power of ChatGPT and combining it with cryptocurrency and blockchain tokens, launching an AI / NFT / crypto platform called KNXT. Let’s unpick the different parts of that picture, because there are certainly a lot of them.

First up, the ChatGPT-derived chatbot, called /madeline. Serving as a bit of a gatekeeper to the “luxury wholesale” site, Madeline (whose name we’re capitalising for the sake of grammar) would have been a deeply fascinating prospect if it wasn’t being introduced in the shadow of a tech landscape that’s been so heavily tilted towards AI all of a sudden that “ChatGPT for X” is now accepted, humorous, shorthand for a very topical goldrush.

But as much as the sheen has rubbed off the power of large language models (at least from the point of view of their novelty), their direct use by brands is still interesting in two ways. 

The first of these is personality: generative models were notoriously easy to coax into going “off model,” and this is something that OpenAI (the company behind ChatGPT) spent a tremendous amount of time guard-railing for its latest version, GPT-4. For a fashion brand, or a group of brands, off-model behaviour represents a tremendous reputational risk. We only need to look at the damage done to the Balenciaga name by a single ad campaign to get an indication of what can happen when a third party is given authority to act on a brand’s behalf – those actions come home to roost quickly.

This may be less of a concern for Kering, however, since the ability to shape ChatGPT (and other LLMs like Claude) into bespoke chatbots with narrowly defined personalities is now something anyone can do, with the help of Quora’s product Poe. Looking to call on ChatGPT via its API and have it act like a luxury fashion salesperson? That’s behaviour you can shape with text prompting alone. And wilder personalities are also just a few keystrokes away.

So, if Madeline isn’t particularly novel as a brand-specific personality layered on top of ChatGPT, what else makes it a worthy case study? The fact that it appears to have been trained on a specific set of products, likely using a vast amount of data from Kering, so it offers pleasing (and accurate) suggestions. The KNXT storefront offers items from Kering-owned companies like Gucci, Bottega Veneta, and Baleciaga.

That type of product dataset is something we’ve talked about several time recently, as part of our analyses of fashion technology news, because for any product discovery engine – chatbots included – to be effective, it needs to be able to index a full catalogue of products, attributes, and use cases in order to make informed recommendations. This is very likely to be the future of brand and retail usage of AI language models – not simply adjusting them to fit their parameters of desirable behaviour, but also making sure that they understand what makes a particular brand’s products unique.

Based on The Interline’s experimentation with Madeline, this training exercise seems to have been quite superficial so far. The model’s recommendations are mostly made on quite soft, subjective metrics, rather than on anything too finite or concrete. And while that’s an acceptable set of criteria for a curated selection of luxury products, a much wider variety of subjective and objective elements will need to be taken into account if a similar approach is going to work for a wider catalogue of different products and brands, or for performance-oriented or sportswear brands where product attributes are more complex and have a greater bearing on customer satisfaction.

image: knxt

In practice, the most pedestrian part of this entire announcement is the crypto component. KNXT puts cryptocurrencies front and centre as the primary method of payment – although the meta description for the portal suggests that card payments are also accepted. Items that are purchased using a crypto wallet unlock an exclusive NFT, which cannot be purchased separately. Those NFTs look to be templated, algorithmically combined profile pictures of the same variety that we’ve seen countless times before, and the open market will bear out whether or not these retain value, or how much benefit they confer in terms of consumer engagement.

Is anything about KNXT unique, then? Not especially. So it’s a valid question to ask whether there is a real forward-thinking strategy behind this kind of platform launch, or whether it’s a case of a well-resourced fashion entity rolling the dice on the latest buzz around AI and NFTs and seeing what happens. 

The strategy question is one that only Kering can answer, but the AI applecart is definitely still rolling in revenue terms. According to new data, spending on the top 10 GPT-3 powered AI apps reached $6 million in February 2023 – a massive 4184% increase since the previous year, as reported by analytics company data.ai. And, crucially, not joining the AI race may be counter-intuitive since a recent survey by Ipsos for the World Economic Forum found that nearly two-thirds of adults expect products and services using AI will profoundly change their daily lives in the next 3-5 years. But what is the longevity of these AI-powered assistants beyond this timeframe? It’s still too early to tell whether the trend will pass or continue to improve and become an integral part of all our shopping experiences, but it seems all but guaranteed that we’ll see similar experiments – and hopefully deeper, more considered ones – from other brands in the very near future.

image: ikea kreativ

But if AI chatbots are one side of tech-enabled product discovery, we also saw their direct opposite hit in the headlines in the home furnishings space this week, with IKEA building a new human-powered interior design service into its already-existing 3D virtual design tool, Kreativ

If you’re not already familiar with it, Kreativ allows you to scan your 3D space and to erase some of your existing furniture, replacing it with suggestions of IKEA’s own furniture. So far this has been a purely customer-directed endeavour – a way for shoppers to visualise products in their own space, with limited recommendations. It would have been almost expected at this point for IKEA to have supplemented that experience with AI recommendations, and while that’s not to say that AI interior stylists aren’t coming, for now at least IKEA is pairing shoppers up with a designer who will map out your space in 3D, create a mood board, and get your input on your specific choices around lighting, materials, and furniture while giving suggestions. Following this, all will be arranged for you in terms of buying and shipping, leaving only payment necessary. 

Two very different ways, then, to bring consumers to what a brand believes is the “right” product for their needs. And yet another stress test for what it might mean for AI to insert itself more meaningfully into product discovery, and how that compares to the human touch.

Emerging digital designers: traditional meets technology 

Earlier this week, digital fashion platform SYKY announced the launch of a new luxury fashion incubator program focused on digital fashion. The announcement follows its most recent funding news in January, where it received just over $9 million to further its mission of enabling creators to curate, generate, share and trade their fashion collections.

It’s encouraging to see yet more support for new and emerging designers, of course, and we’re very clear on the benefits of 3D / DPC for helping new creatives to realise their vision. And it seems that the SYKY Collective initiative has the potential to serve as a launchpad for the blossoming careers of the chosen digital designers. The year-long programme is offering “Web3 mentorship” that will ensure that the emerging designers will have access to industry experts, from traditional fashion brands, who will share their knowledge on how to create and run a successful fashion house.

Some of the leaders to be featured as part of the initiative include Jonathan Bottomley (CMO of Calvin Klein) Megan Kaspar (Creative Editorial Director at Vogue) Megan Kaspar, and Matthew Ball (founding member of Red DAO). 

The designers will also receive access to the British Fashion Council’s curriculum, which has seen notable alumni like Alexander McQueen and Stella McCartney. And it’s not all academic, there will also be an immersion into the fashion industry participants will be granted special access to major industry events, and be able to partner with top luxury brands that the programme will likely introduce them to. This type of pairing between emerging digital designers and established traditional designers could be exactly what the fashion industry needs at this point to take steps toward a future that is cohesive and collaborative with technology and established fashion craftsmanship.

To put on our cynics’ hats for a moment, though, it’s equally possible to look at this news as something more calculating than just an altruistic way to support the future of fashion. The future of digital fashion as an anchor for metaverse and web3 visions is currently looking a little shaky, and for a buzzy new initiative that means finding ways to build its own addressable market and justify its valuation – and onboarding new creators into their ecosystem is certainly one way to do that.

The best from The Interline:

This week, we published a collaboration with Texprocess Americas, and a feature from industry journalist and writer, Emma Feldner-Busztin.

In the run up to the next Texprocess Americas event (May 10 – 12, 2023 at the Georgia World Congress Center in Atlanta, Georgia), we worked with the tea behind the show to unpick why professional development and technology are becoming inseparable in fashion.

Emma Feldner-Busztin, industry journalist, and News and Features Editor for The Interline, asks what the future of real-time rendering could have in store for fashion?