Few industries deal in assets as precisely documented as diamonds. Every stone has a weight, a colour, a clarity grade and often a grading report. Yet much of the trade still runs on fragmented systems, phone calls and manual work, and deals still depend heavily on expertise and long relationships.
That is starting to change, from several directions at once. Tracr, the provenance platform backed by De Beers, has registered more than five million rough diamonds, and in May the Gemological Institute of America agreed to take a 30% stake in it. Sarine, an Israeli diamond technology company, introduced AI colour and clarity grading in 2017 and now offers automated grading of the 4Cs. Rapaport, best known for its weekly price list, is building AI into its trading platform under a new chief executive, Dan Mano, who joined in April 2025 from senior roles at Perion and MyHeritage.
The most obvious place for AI is grading. Sarine’s case is that a machine grades more consistently than a person affected by fatigue or lighting, and that suppliers can grade stones in-house without sending them to an outside lab.
Market data is another. The industry already produces a great deal of it. Rapaport publishes a weekly price benchmark and a daily index built from asking prices on its own marketplace. Mano’s view is that AI can turn data like this into analysis that used to take hours of manual work: what a stone is worth, how competitively it is priced, how demand is moving and how much risk sits in holding it.
Search and matching are harder than they sound. Two stones with identical weight, colour, clarity and grading reports can look different and sell for different prices. Any system that matches buyers and sellers has to account for more than the specification sheet, which is exactly where experienced dealers earn their living.
The more interesting question is where AI should stop. A diamond deal involves authenticity, quality, reputation, money and personal judgement. Mano argues that AI can flag inconsistencies, suggest prices and highlight risks, but should not decide on its own whether someone can be trusted, authenticate a stone conclusively or settle a serious commercial dispute. Measuring a stone is one thing. The judgements around a deal are where he wants people to stay in charge. AI’s biggest value, he says, will come from making experienced professionals better at their jobs.
I think that is the right instinct, and it applies well beyond diamonds. Every traditional industry faces the same question: how to modernise without undermining the expertise and relationships it depends on. In diamonds there is an added reason to get it right. Mining and cutting support communities and livelihoods across Africa and India, and a trade that loses buyers’ confidence puts those at risk.
Mano expects diamond trading to be far more connected and automated within five years, with a retailer able to price, source, verify, pay for and manufacture a piece through a single platform. Whether that happens on one company’s platform or across several, the direction looks clear. The administrative work around a deal will shrink. The people who know diamonds will have more time to spend on the parts of the trade that still need them.
