As the art market continues to debate the broader implications of artificial intelligence (AI), auction houses are witnessing significant advancements thanks to AI-enhanced translation and data processing services.
“Since online platforms have been embraced more readily, there’s been a huge expansion in the international visibility for auction houses, particularly those which were either very regional or domestically focused,” says Julia Marie Wendl, the owner and chief executive of Kunst-Auktionhaus Wendl in Rudolstadt, Germany.
“For a long time, even when we had Google Translate, we were relying on staff to work through translated lot details to ensure accuracy. With the advance of AI’s ability to train itself, we are now able to help the system recognise and implement sensitivities around art historical or market terminology,” Wendl adds.
Antoine de Rochefort, the head of Drouot.com, similarly emphasises how crucial AI has become for handling large quantities of data: “Our challenge differs from that of traditional auction houses due to the large volume of items we offer for sale [three million annually], from 790 different auction houses, each providing non-standardised information,” he says. AI services not only help with the translation of material, but also with the automated categorisation of items from Drouot.com’s network of partner vendors, he adds.
Tackling the challenges of e-commerce
For platforms offering a sizable array of items, quickly directing users to products and providing content in their preferred language are key to business. “These are challenges that e-commerce has been able to adeptly manage for quite some time now, though in retail, products are stocked and available for purchase on a long-term basis,” Rochefort says. “This is not the case with auctions, where products are unique and will remain on the market for a very short time. Traditional e-commerce selling practices are therefore not applicable in this case.”
Tackling such challenges will be worth the effort. At the recent Art Business Conference in London, Stephan Ludwig, the chief executive of Forum Auctions, said: “A recent study suggested that only 14% of luxury furnishings purchases are made online. I gather it’s about a £10bn a year market in the UK. The notion of us being able to convert, as an industry, 2% or 3% of that spending into our industry would probably double the size of the regional UK market.” He continued: “Sustaining relevance for users and furnishing an e-commerce experience for auctions necessitates a hyper-personalised approach, one that only AI can systematise industrially.”
Human involvement
As ever, opportunities presented by AI are being adjoined with assurances that human involvement is still crucial. “It’s not just the language sensitivities; it’s cultural differences, too,” Wendl says. “For example, in Germany, the price we provide is usually a starting point to negotiations, not fixed. This has to be explained and discussed with clients.”
Peter Greenway, of the London auction house Roseberys, confirms that AI’s lack of nuance still means “relying on our in-house team, fluent in over 25 languages, to handle these communications. In high-stakes international sales, trust and precision are paramount, and many of our clients prefer the personal, human-led approach for these reasons”.