One of the most interesting discussions at yesterday’s Art Business Conference was about what AI is doing behind the scenes and its transformative potential for the art world, writes Charlotte Heath-Bullock.

The examples were striking. Edouard Gouin of Convelio, who's been on a decade-long mission to simplify art logistics, described AI agents that can process quote requests, flag works requiring export licences and qualify incoming sales leads. One quoting agent alone is reportedly saving his team around 25% of their time.

Sam Glattman of ArtSignal talked about AI-assisted valuations capable of assessing huge numbers of individual objects while taking account of variables such as condition and provenance.

And Brendan Beu of Synergy Associates described AI trawling the mass of unstructured information sitting within emails, WhatsApp messages and databases to identify connections between available works and potential collectors, opportunities that might otherwise never surface.

The immediate temptation is to frame all of this as a story about automation and efficiency. But perhaps the more interesting consequence is what happens to the humans.

The art market remains an intensely relationship-driven business. Collectors want to talk to specialists and clients value judgement, discretion, knowledge and trust. As the panel observed, nobody entered the fine-art world because they wanted to spend their days doing administration.

If AI can take away some of the quoting, form filling, searching, checking and sorting, it potentially gives people more time for the work only people can do: advising clients, developing relationships, applying judgement, finding opportunities and building trust.

There are important caveats. AI is only as useful as the information sitting beneath it. Much of the art world still operates with fragmented databases, legacy systems and decades of valuable knowledge buried in inboxes and spreadsheets.  

And cybersecurity may become the biggest challenge of all. AI doesn’t just make businesses smarter; it makes attackers smarter too. The discussion around prompt injection, autonomous attacks and the vulnerability of sensitive client information was a useful reminder that enthusiasm for AI needs to be accompanied by considerably greater technological vigilance.

So perhaps the businesses best placed to benefit from AI over the next three years won’t simply be those that adopt it fastest. They will be those that first get the fundamentals right: clean data, robust systems, clear processes and strong cybersecurity.

The irony is that if they do, one of AI’s greatest contributions to the art business may be allowing an increasingly digital industry to spend rather more time being human.

Find out more here.