Rapaport Research Report: Rise of the Machines
Imagine the following: A diamond-mining company operating with self-driving trucks loads its haul of ore to the processing plant, where rough diamonds are separated from waste and arranged in their rough forms into different sizes before being sent for further aggregation. The goods are fed into machines that determine their color and clarity potential while also assessing where the stone fits best amid the many assortment options. The diamonds are then run through the company’s computer-based planning mechanism, which analyzes how to get the optimal polished yield.
From there, the stones go to the polishing department, where machines use artificial intelligence (AI) to carry out the cutting and polishing process. The resulting diamond receives a grade via automated systems covering the four Cs. It then gets shipped to the jeweler or end-consumer, and all its information is uploaded to a blockchain or cloud-based traceability program.
That process was presented as a real possibility in a panel discussion at the Dubai Diamond Conference in September, which revolved around the theme of “Disruption in Diamonds.” The session posed the question: How will increased manufacturing automation affect our supply chain? Moderator Anish Aggarwal, founder of consultancy Gemdax, probed further: What will automation mean for the hundreds of thousands of workers employed in the manufacturing sector? And how long before we get to full automation of the manufacturing process?
The answer, the panelists agreed, is sooner than you think.
“There is pressure in the market and strong demand for a solution,” stated Bernold Richerzhagen, founder and CEO of Synova, which that same week launched DaVinci — an automated cutting and shaping system for diamonds. “We’re ready to propose the solution, and it’s a question of how long it will take to implement.”
Diamonds in a day
The push toward automation stems from the lack of efficiency in diamond distribution channels, asserted David Block, CEO of Sarine technologies, which provides equipment used in diamond manufacturing. The fact that there are over 100 steps in the manufacturing process is a case in point, he added.
The trade’s inability to manufacture according to consumer demand has resulted in the excess stock of a year or two being stuck in the pipeline, Block continued. While it can take about a year for a newly mined rough diamond to make it to the market at present, technology can shorten the cycle by helping the trade adapt the rough to what the consumer wants, he said.
Speeding up the manufacturing process is an opportunity that De Beers recognized when developing its innovation strategy. Looking at ways to make the pipeline more efficient, the company found that the biggest bottleneck between miner and consumer lay in cutting and polishing, recalled Faried Sallie, head of technology at De Beers.
To condense the pipeline, “we were told our ambition should be to mine the product and get it to the customer in a way that it will be ready in one day,” he recounted on the panel. Grappling with that challenge led De Beers in 2015 to acquire a 33.4% stake in Synova, since the Swiss company’s microjet technology stood the best chance of being adopted by the industry as a long-term solution, Sallie said.
Revolutionizing the system
Four years later, Synova claims its DaVinci system can drastically reduce the polished-diamond production time by combining several manufacturing processes in one machine. Using waterjet-guided laser technology and a computer numerical control (CNC) system that processes complex three-dimensional geometries, DaVinci can tell when a cut is finished and then automatically initiate a change to another facet on the diamond.
Synova is not the first company to announce a tool that automates polishing. Last year, the Antwerp-based Scientific and Technical Research Center for Diamond (WTOCD) unveiled its Fenix equipment, which the center’s researchers predicted would make the manufacturing process 10 to 20 times faster.
Both companies claim their technology will fundamentally change the industry and reduce dependence on high-cost labor.
“It will revolutionize the diamond manufacturing industry because the [DaVinci] system covers virtually the complete rough-to-polished process,” Richerzhagen noted at the launch. “Several cost-, skill- and labor-intensive steps in the polishing phase — such as the crown and pavilion blocking, girdle bruting or recurrent quality checks — become redundant.”
The consequence is that manufacturing can be adapted to seasonal demand and location, Synova added, and the greater accuracy and improved stone symmetry mean customers gain a higher and more predictable polished yield.
All of that has implications for the manufacturing workforce, as the panelists and many concerned members of the trade pointed out at the conference. Sallie and Block drew on their respective experiences to put the potential labor impact of automation into context.
In the past, De Beers has deployed driverless trucks for certain parts of the mining production process. Today the company is looking at autonomous loading, hauling and drilling machines that make decisions with no human intervention, Sallie reported.
The result, he added, is a safer and more efficient process, and that enables the industry to tell a better story.
At the same time, De Beers has engaged with communities and governments to find alternative, value-adding jobs for the workers affected by these new systems, Sallie went on. However, he acknowledged that it wasn’t always possible. “There’s inevitably going to be times when you either have to retrain people or let them go. Our partners recognize that to be competitive, there is going to be this tension.”
Concern about the effect innovation might have on labor is not a new phenomenon, nor is it unique to the diamond industry, Block noted. He pointed to the industrial revolution in the 1800s, which many said would take away jobs. But the net effect has been positive, he stressed.
“Jobs evolve, and sometimes they disappear,” he said. “But in the big picture, when you look at how technology has added value to the industry, on average it doesn’t take away jobs, it changes them.”
He recalled the industry’s resistance to Sarine’s rough-planning machines when they first came onto the market about 20 years ago. Diamond planners were worried about their jobs, he said, but Sarine opened a school to train them in using the technology, and today there are thousands working in diamond planning. The key, he stressed, is to show people how to utilize the technology at an early stage of the process.
A wider scope
What is certain, Block declared, is that change is inevitable.
As the technology evolves and improves, systems such as DaVinci and Fenix will be able to manufacture a wider range of diamonds, as the trade’s experience in other areas suggests. Sarine’s rough-planning machines, for example, started with bigger and better-quality diamonds, but today, they’re working with stones as small as 0.01 carats as well, according to Block.
Similarly, in diamond grading, technology is enabling labs to address diamonds in sizes that were not possible to assess in the past, noted Tom Moses, executive vice president and chief laboratory and research officer at the Gemological Institute of America (GIA).
Panelists agreed that over time, the diamonds being manufactured by automated systems would include progressively smaller and cheaper stones.
While the executives varied in their estimations of how long it would take the industry to become fully automated, they noted that different processes were making use of new technologies all the time, thereby speeding up the journey. For instance, Sallie pointed out, components such as financial systems and human resources are already being automated.
So, asked Aggarwal, where does that leave manufacturing companies?
Manufacturers need to prepare themselves before the change happens, Block replied. Sallie concurred, saying they would have to be open to the new possibilities that automation, AI and other technologies presented if they wanted to better meet the needs of their customers and the end consumer.
The adoption of technology means the industry is going to compete in a different way, the De Beers executive predicted.
“What we do today is not what we’re going to be doing tomorrow, and we’re not going to be selling the same product. It’s not about replacing our existing manufacturing capability. We’re going to [apply] our offering in very different ways, which will open the possibility to reframe our product.”
Machine vs. Gemologist
There’s a debate in diamond-grading circles regarding the future of gemologists. Some, most notably Sarine Technologies, believe the grading process will become fully automated as technology evolves. The Gemological Institute of America (GIA), meanwhile, expects that gemologists will always be necessary to verify the machines’ findings.
The GIA developed its first color-grading machine as far back as 1945, according to the institute’s Tom Moses. More recently, the group has invested in machine learning and artificial intelligence (AI), though its aim is to improve efficiency rather than replace existing graders, he emphasized during a panel discussion at the Dubai Diamond Conference in September.
“We see this move toward automation to handle more stones and at a price and efficiency that supports consumer confidence,” Moses said. “But I don’t think we’ll get to a point where grading will become wholly automated without human graders.”
In contrast, Sarine predicts just such an eventuality. Sarine has devices that it claims can automate the grading of all four Cs, with color and clarity being the most challenging. It’s all based on AI and machine learning: The equipment gathers data from repeated tests and analysis of tens of thousands of diamonds.
Implementing a fully automated grading process reduces the chance of human error, enhances consistency, and makes it less subjective, contended Sarine CEO David Block, who has described Sarine as a technology company supported by gemologists.
“Is a GIA F-VS2 the same today as it was 20 years ago?” he challenged the panelists. “It’s an issue of bringing consistency over a long period of time. It’s inevitable that manufacturing will evolve toward automation, and the same applies to diamond grading.”
The GIA is starting to phase in the use of AI in its color and clarity grading, and technology is also enabling its labs to assess diamonds in smaller sizes, which was not possible before, Moses said — though he added that these applications would still require verification from a gemologist.
If grading does become more automated, asked panel moderator Anish Aggarwal, wouldn’t manufacturers just use the machines on their own, rather than send stones to a lab?
That is a possibility the GIA is considering, Moses acknowledged.
“We do anticipate a scenario in which at least for some diamonds, we would rely on technology for a grade, with the results being collaborated and monitored by a gemologist,” he said. The data would be collected, then uploaded to the cloud and verified on the GIA’s system, a process he said could work for a substantial part of the production stream.
This article was first published in the October issue of Rapaport Research Report.
Image: Panelists in the discussion on technology disruption at the Dubai Diamond
Conference (from left): Anish Aggarwal, Tom Moses, David Block, Faried Sallie,
Bernold Richerzhagen. (DMCC)