At IWC Schaffhausen’s booth at the Watches and Wonders fair in Geneva earlier this year, Maurice Moitroux, IWC’s associate director of brand marketing, gestured towards a large, clunky, beige device sitting atop a counter near the entrance.
“It looks like a microwave, but it’s actually a computer,” he said.
Intended to conjure the retro-futuristic vibe of IWC’s new Ingenieur watch collection — an update to a 1976 model famously designed by Gerald Genta — the vintage NCR monitor, dating to the 1980s, was one of several period artefacts displayed in and around the booth, including an experimental Mercedes-Benz C 111-III car from the 1970s.
The computer, however, contained something unexpected: a ChatGPT 3.5-series module connected to a voice-activated application built to answer questions about Genta, who died in 2011, leaving behind a well-documented horological legacy.
The display was an apt metaphor for the luxury watch industry’s current relationship with artificial intelligence. On the surface, AI is still something of a non-starter, with few brand executives being willing to discuss it publicly. (A group of tech industry leaders, however, recently signed an open letter warning of the existential threat it may pose to humanity.)
Behind the scenes at big brands such as Cartier and Panerai, however, machine learning, a subset of AI that involves training machines to learn from data, is quietly revolutionising how watches are made and brought to market.
A handful of watchmakers have admitted to using ChatGPT to generate snappy social media copy and the AI-based design platform Midjourney to whip up wacky mood boards (as in, “We want to do a shop and have it to look like it’s underwater in the Maldives,” George Bamford, a London-based watch customiser, said on a recent call).
But very few are communicating openly about it.
“Today, the watch industry is selling a high-end luxury experience,” Serge Maillard, publisher of the trade magazine Europa Star, said by phone from Geneva. “If AI can be used to elevate the experience, then the term AI might be a bit more present in the speech. But as long as it remains something more in the background, it will be in the shadows.”
Maillard said one reason the industry hadn’t mustered more enthusiasm for AI than it had for technological novelties such as NFTs, blockchain and the metaverse might have to do with the collapse last year of the NFT and cryptocurrency markets, which left them wary of unproven technologies.
“AI is a more silent revolution than NFTs but a deeper revolution,” he said. “It’s something everyone might use one day.”
For the handful of executives who have adopted it, AI has proved transformational.
“It’s saving me time with my design team on research, mood boarding,” Bamford, the founder of Bamford Watch Department, said during a recent phone call. “I now freely admit that instead of going on Pinterest and Google to look up images, I’m putting in wording in Midjourney that will create a unique image. It doesn’t inform the design, but it starts the narrative.
“I wouldn’t use it to design a watch,” he said. “I’ve tried it a few times, and it always comes back like something you don’t want to see.”
Benjamin Arabov, the chief executive of Jacob & Co., said he had a similarly disappointing design experience with A.I. when he used text prompts to generate a watch image. The concepts were impossible to manufacture, “like a waterfall flying in the watch”, he said in March at a company event in Geneva.
Arabov has, however, found AI useful for copywriting, especially for social media. “ChatGPT has become a really good resource,” he said. “‘Here’s all this info about this watch. Now, elegantly explain this timepiece in 100 words’ — because no one is going to read past that.”
Algorithms at work
In the manufacturing realm, watchmakers have used machinery powered by algorithms for at least the past five years. At Roger Dubuis, a system called Jarvis, named after the loyal butler in Marvel Comics’ “Iron Man”, “helps regulate machines, telling us if they are not so precise anymore, and they talk to each other,” Nicola Andreatta, the chief executive of Roger Dubuis, said at Watches and Wonders Geneva.
Panerai, which, like Roger Dubuis, is owned by Richemont, employs a similar system that has been in place for about two years, said Jérôme Cavadini, the brand’s chief operating officer.
“We connect machines to machines, and they talk together,” Cavadini said at Watches and Wonders. “We are able to detect at what moment we are supposed to change the tools, and we can adjust their speed, detect temperature deltas and act accordingly to modify the digital settings of
Where AI, or machine learning, is poised to make the biggest impact on the watch industry is in helping to anticipate demand and speed goods to market.
“We developed algorithms and use them on a monthly basis to detect sales trends, and especially the deviation to the standard forecasts that we do,” Cavadini said. “The point is to connect demand to supply.”
The demand planning system, which Panerai implemented last September, helps the brand detect which products are selling out in certain markets so it can optimise their distribution. “Produce the right reference, for the right market, at the right moment,” Cavadini wrote in a follow-up email.
During a press roundtable event at the Cartier booth at Watches and Wonders, Cyrille Vigneron, the brand’s chief executive, said that AI was also transforming that brand’s demand planning and deployment models — and even giving Cartier insights into customer behaviours and attitudes.
“By having so much data coming from call centres or web inquiries or even what words customers use on comments or on their own social media, you have a lot of words,” Vigneron said. “By trying to identify these words or some meanings behind them, it’s the opposite of ChatGPT — you don’t try to create content, you try to understand what it means. For example, you have a new product coming, and you can have a big hike on social media, which means it’s maybe a good time to increase your production before it has been visible in the sales.”
Vigneron emphasised that AI would not alter Cartier’s design approach, which he described as entirely offer-driven.
“When we design something old but very new again like the Tank Normale, we don’t know if it will meet the public, and that’s part of the beauty in this sector,” he said. “We’re not trying to make an algorithm knowing what customers would want, but to make things we think are beautiful and hoping we find a public for them.
“What we’ve done for the last five years is be distinctively ourselves,” said Vigneron. “That’s collective intelligence, and not artificial.”
Brands aren’t the only watch-world entities looking to optimise their operations with AI.
Wristcheck, a Hong Kong retailer of pre-owned watches, is planning to introduce what it calls Wristcheck Intelligence this year. It is market-analytics software powered by machine learning that “comprehends intricate data patterns to curate an index and suggest ideal prices for our sellers,” Austen Chu, the company’s founder, wrote in an email.
“This software, a product of our proprietary in-house development, has been on our blueprint since the early part of this year,” he wrote, adding that it “exemplifies how AI can facilitate a deeper understanding of market conditions”.
Despite the efficiency-boosting potential of AI, however, many, if not most, watchmakers remain sceptical.
“I’m not against it because you have to be curious and in sync with one’s time,” Philippe Delhotal, the creative director of Hermès’s watchmaking division, said. “But I think the human being has this capacity to bring warmth, doubt, uncertainty, what artificial intelligence does not have. Artificial intelligence is cold, it’s calculated. I think the human margin of error is what gives some things their charm.”
Yet some AI advocates are eager to minimise the number of humans they need to interact with in the course of making and selling their watches.
Ben Waite is one. As the watch designer at his family business, Titan Black, a London company that specialised in customising timepieces, Waite grew tired of working with so many different artisans, he said on a recent video call. The time-consuming process was also complicated by the fact that the original manufacturer’s warranty on the watches he personalised was voided by his work, he added.
Titan Black stopped taking orders about two years ago, and about a year later, Waite fell down the AI rabbit hole. Now, he’s preparing to introduce a company that uses a combination of AI software and a 3D product configurator to design watches featuring bespoke details and to view them from multiple angles, speeding the custom-design process.
“My first job at Titan Black was to design watches and now, effectively, I’ve put myself out of a job,” Waite said. “Now I can generate 20 to 30 designs straight away. The software is designed to help me serve more customers and get to the point quicker.”
“I’m trying to build a whole company with as few people as possible with everything run on AI,” Waite added. “That’s my dream: to have to deal with as few humans as possible. I’m very good at making things but not very good at managing people.”