Sotheby's Buys Machine Learning Firm to Discern the Art You Love

(Bloomberg) -- Who needs an eye to find the next Picasso when AI can do it for you?

That’s what Sotheby’s is banking on with the purchase of Thread Genius, an artificial intelligence startup that specializes in image recognition and recommendation technologies.

The acquisition is Sotheby’s latest step to innovate in the hidebound auction business which has been selling art the same way for centuries. Sotheby’s hopes to use Thread Genius’ technology to match buyers with available artworks based on their tastes.

“Historically, transactions would come together by matching an object with an individual’s preference at a certain price point, and Sotheby’s retains data in those three areas,” Tad Smith, Sotheby’s chief executive, said in a statement on Thursday. “By using advanced technology, we can deliver better insights and match those three elements faster, providing our staff with better tools, our clients with a better experience, and our consignors with even better results.”

Taste Matters

Thread Genius, based in New York, was founded in 2015 by engineers Andrew Shum and Ahmad Qamar. They’ll join Sotheby’s this week. They previously worked at the Spotify streaming service where they helped develop the technology behind its music recommendations. At their own firm they applied a similar approach to the fashion industry, using AI to understand customer tastes based on visual recognition.

“Thread Genius was founded on this principle that no amount of words can fully articulate our taste,” according to the startup’s page on Crunchbase.com. “This is why when we’re asked about our taste in music, we play songs. For fashion, we show pictures.”

While the art world has been slow to embrace technologies that are transforming industries from finance to retail, Sotheby’s is taking such steps. In the last three years, it has hired several finance executives to bolster sales and branding efforts. More recently, it began beefing up the ranks of data specialists.

Richard Vibert joined Sotheby’s last fall as head of data and analytics strategy. He’s organizing the information the auction house collects from sales and consignments and developing methods to identify customer preferences. Vibert joined from the Hong Kong venture capital firm Arbor Ventures, where he was an associate specializing in investments in machine learning and blockchain.

Shum was formerly a software engineer at Spotify. He received his S.B. in electrical engineering and computer science, with a specialty in physics, from the Massachusetts Institute of Technology. Qamar was a machine learning engineer at Spotify and received a dual B.A. in mathematics and physics from the University of Chicago.

The deal builds on Sotheby’s 2016 acquisition of the Mei Moses Art Indices, a database of almost 50,000 repeat auction sales in eight collecting categories. Terms of the Thread Genius transaction weren’t disclosed.

To contact the reporter on this story: Katya Kazakina in New York at kkazakina@bloomberg.net.

To contact the editors responsible for this story: Margaret Collins at mcollins45@bloomberg.net, Alan Mirabella, Josh Friedman

©2018 Bloomberg L.P.