AI is here, and it is not going away. Whether you sell crystals from a folding table at a local gem show or import containers of rock from the other side of the planet, it is going to affect the way you do business. The only real question is whether you learn to use it before the people around you do.
Before getting into that, it is worth clearing up what most people mean when they say “AI,” because the term has been wrapped in so much science-fiction language that it is easy to imagine something far stranger than what is actually sitting on your phone or computer. Most of the tools people are using today are built around large language models and increasingly multimodal systems that can work with text, images, spreadsheets, files and other kinds of information. They are not self-aware, they do not have intentions, and they are not sitting there plotting to replace your rock shop. They are extremely sophisticated pattern-recognition and information-processing systems.
Calling them an advanced version of Google is not technically perfect, but it is a useful starting point. Traditional search engines give you places to go look for an answer. A large language model can take your question, combine several pieces of information, recognize patterns between them and give you something much closer to a finished answer, recommendation or piece of work. That difference is beginning to matter enormously in the rock, mineral, fossil and gemstone industry, because it is changing not only how we identify and sell stones, but also who has leverage in the business.
For decades, one of the greatest advantages in the rock world was capital. The dealer with more money could travel farther, buy more inventory, negotiate larger deals, carry deeper stock and survive the slow periods. That is still true. What is changing is that capital is now being joined by another kind of advantage: technological leverage. Dealers who know how to use AI, e-commerce, analytics and automation are able to learn faster, move faster and organize their businesses at a pace that was simply not possible before.
Stone Identification Has Changed
One of the most obvious applications is stone identification. For years, companies have built apps specifically to identify rocks and minerals from photographs, and in some cases entire businesses were built around that one function. Now I can take a photograph of a specimen, upload it into an AI system, and add whatever context I have: where I bought it, what country the dealer said it came from, its hardness, whether it reacts with acid, how heavy it feels, what other minerals occur with it, or even just where I found it.
That context makes a major difference. A photograph alone can be misleading because calcite, quartz, feldspar, fluorite and dozens of other minerals can occur in enough colours and habits to confuse visual identification. But if you combine several photographs with locality information and some basic physical properties, you suddenly have a surprisingly competent identification assistant sitting in your pocket.
That does not mean AI has replaced mineralogical testing, and it would be reckless to pretend otherwise. If I am buying an expensive gemstone and the difference between natural, synthetic, heated, filled, dyed or untreated material represents thousands of dollars, I am not betting the farm on a chatbot looking at a photograph. Refractometers, microscopes, spectroscopy, specific gravity, laboratory reports and experienced human eyes still matter. What has changed is that the early stage of the investigation can now happen much faster, and for a working rock dealer handling hundreds or thousands of different materials, that speed has real value.
The Encyclopedia Behind the Counter Is Disappearing
There was a time when almost every rock shop had some kind of encyclopedia behind the counter. Sometimes it was mineralogical, sometimes it was a gemstone reference book, and sometimes it was one of those enormous metaphysical crystal guides explaining what amethyst was apparently doing for your third eye on a Tuesday.
A customer would walk up with a stone and ask what it was. The employee did not know, so out came the book. They would flip through photographs, check the index, read a paragraph, and do their best to answer the question.
Today, the employee can photograph the mineral, describe what they know about it, and get a starting point almost instantly. From there, they can ask follow-up questions: where does this material commonly come from, what minerals usually occur with it, how does it form, what is the difference between it and a similar-looking mineral, is it commonly dyed, why is this specimen this colour, and what should I tell a beginner who knows absolutely nothing about geology?
That does not make the employee a geologist. It gives the employee access to a very capable research assistant, which is only useful if the person using it still knows enough to recognize when that assistant is confidently talking out of its ass.
AI Is Speeding Up the Crystal Product Pipeline
The bigger impact for dealers may not even be identification. It is everything that happens after we buy the rock.
Think about what has to happen before a single specimen becomes a sellable online product. Someone identifies it, photographs it, edits the photograph, measures it, writes a title, writes a description, decides what collection it belongs in, creates tags, thinks about search terms, uploads it to the website, and may also need to create social-media copy for it.
Now multiply that by 500 pieces, or 5,000.
This has always been one of the great difficulties of running a serious rock business online. Rocks are horrible inventory. If you sell toasters, you photograph one toaster and sell 20,000 identical units. If I buy 300 mineral specimens, I may have 300 individual products, and every one of them has a slightly different crystal formation, weight, colour, dimensions, condition, value and photography requirement.
AI does not eliminate that work, but it changes how quickly a competent team can move through it. A knowledgeable employee can provide the facts and photographs while AI assists with descriptions, formatting, tagging, categorization, SEO structure and repetitive administrative work. Image tools can increasingly help with backgrounds and presentation. Product information can be reformatted for different platforms, and data-entry workflows can be standardized instead of rebuilt from scratch every time.
The interesting part is that this does not necessarily mean the business spends less money. In my case, I am not spending less because of AI. I am doing more. Instead of using technology to eliminate the photography and data-entry budget, I can use the same people and resources to put more inventory online and improve the quality of the work around it. The productivity ceiling moves, and that distinction gets missed constantly when people talk about automation. Sometimes automation does not shrink the business at all. It simply gives the business more capacity.
Crystal Customers Now Have Better Information Too
Dealers also need to understand that this works in both directions. Customers now have access to much better information than they did even a few years ago, and that changes the relationship between buyer and seller.
Someone considering a $700 mineral specimen can photograph it and ask an AI system what it appears to be, whether material from that locality commonly sells in that range, what affects the value, what damage to look for and what questions they should ask the dealer before buying. The answer may not be perfect, but the information imbalance between dealer and customer is shrinking.
For decades, experienced dealers had a major advantage simply because they knew things the customer did not. They knew where the material came from, which stones were truly scarce and which were only marketed that way, what comparable pieces traded for, which treatments were common and what warning signs to look for. That expertise still matters enormously, but simply possessing obscure information is becoming less defensible as a competitive advantage. The businesses that survive that transition will probably be the ones that genuinely know their products rather than relying on the fact that their customers do not.
I consider that a healthy change.
Where AI Gets Much More Powerful: Running the Crystal Business
The part I find more interesting than stone identification is what happens when AI starts sitting on top of the business itself.
A modern rock shop can have information scattered across Shopify, Google Analytics, Google Ads, Meta, Instagram, email, accounting software, spreadsheets, inventory records and customer databases. Historically, actually understanding all of that required either a fairly sophisticated owner or several different specialists. Even then, everyone tended to look at their own little island.
The marketer looked at advertising. The bookkeeper looked at accounting. The retail manager looked at the store. The website person looked at traffic. The owner tried to figure out whether any of it connected.
That is beginning to change.
I can now look at what products actually sold during a month, compare that with search behaviour, examine advertising performance, look at customer acquisition, determine which categories are growing, examine average transaction values and use all of that information when deciding what the marketing department should prioritize next month. Instead of looking at ten isolated reports, I can increasingly ask how the pieces relate to one another and where the business is actually moving.
That is considerably different from using ChatGPT to write an Instagram caption.
AI Is Becoming a Management Tool
This is the part I did not expect when I first started using it seriously. AI is becoming useful not just for marketing or content, but for operations and management.
We have standard operating procedures for different parts of the company, and those procedures can be reorganized, updated and converted into training material quickly. Complicated processes can be turned into flowcharts. Employee training documents can be compared against job descriptions. Responsibilities can be separated between positions, and recurring problems can be examined from multiple angles.
If something repeatedly goes wrong inside the business, I can work backwards through the problem and ask whether the failure is coming from the employee, the training, the incentives, the instructions or the system itself. That kind of analysis is not magic, but it is much easier when you can put the whole problem in one place and interrogate it from several directions.
The same thing applies to reporting. Key performance indicators can be pulled together monthly and examined as a system rather than one by one. Different versions of the same information can then be generated for different roles inside the company, because the owner needs one kind of summary, the marketing department needs another, the retail supervisor needs another, and the bookkeeper needs something else again.
More importantly, those reports can contain priorities rather than simply numbers. The machine is not running the company, and it should not be. The owner still has to decide what matters, judge whether the recommendations make sense and understand the consequences of the decisions. What has changed is the amount of information one competent person can process and organize without needing an entire management team around them.
That is where AI stops being a novelty and becomes leverage.
The Old Advantage in Crystals Was Capital
The rock industry has an unusual demographic structure, particularly in North America, where a large amount of established inventory and purchasing power sits with older dealers who built their businesses over decades.
There is a perfectly logical reason for that. Rocks require capital.
You need money to buy inventory, then more money to buy larger quantities. Eventually you are buying pallets, crates and containers. Maybe you are travelling overseas, maybe you work directly with factories, or maybe you are having spheres, towers, carvings or cabochons manufactured for you. Over time, the scale itself becomes an advantage because the dealer with more capital can buy larger quantities at better prices, build stronger supplier relationships, carry deeper inventory, negotiate harder and survive slow periods more comfortably.
A 23-year-old selling crystals on Instagram might understand social media brilliantly, but if her entire inventory is worth $20,000 and another dealer can spend $20,000 during breakfast at a buying show, they are not playing the same game.
Capital compounds, and for decades that advantage helped shape the mineral industry.
It still matters. The difference is that technology is now beginning to compound alongside it.
Technology Is Becoming Leverage in the Crystal Business
For the first time, very small companies have access to a level of organizational and analytical capability that once required substantially more people.
A technologically competent entrepreneur can learn SEO, analyze advertising, build landing pages, examine customer behaviour, create email campaigns, improve product photography, study competitors, build employee training systems, improve inventory workflows, research pricing, learn management systems and communicate directly with suppliers around the world. When they do not understand something, they can sit with an AI system for three hours and keep asking questions without exhausting a consultant or waiting for someone to get back to them.
That does not replace experience, and it certainly does not replace money. What it does is allow somebody who is ambitious, technologically literate and willing to work to climb certain learning curves dramatically faster than previous generations could.
That changes competition because it changes the speed at which a business can improve.
COVID Already Showed Us What Happens When Distribution Changes
The rock industry had already started moving online before AI became mainstream, and COVID accelerated that transition dramatically. Live sales exploded, Instagram selling became normal, and WhatsApp and WeChat became major parts of international sourcing. Dealers in Pakistan, India, China, Indonesia, Morocco and elsewhere could reach North American buyers directly without necessarily waiting for them to walk around a gem show with a badge around their neck.
Collectors became more comfortable buying specimens from photographs and videos. Factories gained direct access to retailers, retailers gained direct access to factories, and collectors gained direct access to wholesalers. Everybody moved one step closer to everybody else.
The traditional gem show is not disappearing tomorrow, and serious buyers still know there is no substitute for handling material in person. Shows remain incredibly important for relationships, discovery, negotiation and the kind of accidental finds that do not happen when you are searching a website by keyword. What has changed is that they no longer have a monopoly on access.
A dealer who once depended almost entirely on physical traffic is now competing against businesses that know when their customers are online, what products they interact with, which advertisements brought them in, which emails convert, what videos they watch and, increasingly, what those customers bought six months ago. Once AI is added to that pile of information, the advantage is no longer simply that one dealer is “better at social media.” The entire business becomes more measurable.
Capital Leverage Is Meeting Technology Leverage
This is where I think the next interesting battle in the rock industry happens.
For decades, established dealers held an enormous advantage because they had capital, inventory and supplier relationships that took years to build. Many younger sellers now possess a different kind of leverage: they are faster with technology, more comfortable selling online and more willing to build their businesses around data, automation and digital communication.
Neither advantage automatically wins. There are older dealers who are exceptionally good with technology, and there are plenty of young crystal sellers whose entire digital strategy still consists of pointing an iPhone at a table and yelling “CLAIM” into Instagram Live. The real divide is not age. It is adaptability.
The businesses that will be hardest to compete with are the ones that eventually combine both advantages. A dealer who understands minerals, understands international buying, has enough capital to purchase serious inventory and also understands modern e-commerce, analytics, automation, content and AI is operating with a completely different set of tools than someone who only has one side of that equation.
The old dealer might have three million dollars worth of rocks but almost no digital infrastructure. The new dealer might have extraordinary digital infrastructure but only thirty thousand dollars worth of inventory. Neither is unbeatable.
Eventually, somebody is going to have both.
The Next Generation Can Build the Crystal Business Faster
This is the part established dealers should pay attention to.
Previous generations accumulated knowledge slowly because there was no alternative. You learned by working shows, getting burned on purchases, developing relationships over decades, experimenting with advertising, hiring consultants, buying books and talking to other dealers. You spent years discovering which systems worked inside your business because there was no machine sitting beside you capable of helping you organize the lessons.
Experience still matters, and there is no shortcut for actually handling millions of dollars worth of stone over a lifetime. There is no app that can give you twenty-five years of buying experience overnight.
What has changed is everything surrounding that experience.
A younger entrepreneur can make ten mistakes, feed those mistakes into an AI system, analyze what happened, research better approaches and redesign the process far faster than someone could have done twenty years ago. They can borrow systems from industries completely outside the rock business, adapt those systems to their own operation, test them, measure the results and adjust again.
Business is, in large part, an iteration game. The company that can learn faster, test faster and correct mistakes faster has an obvious advantage over the company that relies entirely on habit.
That may be the biggest advantage AI gives us. Not answers, but speed of learning.
AI Does Not Replace the Rock Dealer
There is an important distinction here because this is where people tend to overstate what the technology can do.
AI cannot walk through a warehouse in Brazil and develop the instinct that tells an experienced buyer which pile has the good material. It cannot pick up a piece of jade and feel the density, notice that a dealer is behaving strangely, or understand that a parcel looks suspicious because twenty years of experience is telling you something is wrong. It does not spend decades developing trust with a family that owns a mine, and it may not notice the tiny repaired crystal halfway down a flat unless somebody shows it the right photograph.
It does not have taste, reputation or relationships, and it does not have judgment in the human sense.
Those things remain incredibly valuable.
The mistake is assuming that because AI cannot replace those parts of the business, it is therefore not going to change the industry. It will, because the dealer still supplies the judgment, experience, capital and relationships while the machine increasingly amplifies everything around them.
That combination is what matters.
The Crystal Dealers Who Refuse to Learn AI Are Giving Away an Advantage
I do not think everyone in the crystal business needs to become a programmer, and I do not think you need to understand neural-network architecture or how a transformer model predicts tokens. You need to understand what these tools can actually do for your business.
Start with something boring. Take an SOP you hate updating and have AI reorganize it. Photograph a mineral you are uncertain about and investigate it. Give it a sales spreadsheet and ask what patterns you might be missing. Have it explain your Google Ads report. Use it to help create the first draft of fifty product descriptions. Ask it to analyze the questions customers repeatedly ask your staff. Feed it your monthly numbers and build a KPI report, then take that report and ask what questions you should be asking that you are not.
The important skill is not memorizing every AI tool that exists today, because half of them will probably be irrelevant a few years from now anyway. The skill is learning how to work with systems like this and how to integrate them into the way you think, manage and operate.
I think the competitive divide forming in the mineral industry is much bigger than people who use ChatGPT versus people who do not. It is businesses that are increasing their rate of learning versus businesses that are not.
For most of the history of the rock trade, money, inventory and relationships created enormous barriers to entry. Those barriers still exist, but technology is giving smaller and younger companies a new kind of leverage, and that leverage will only become more powerful as the tools improve.
That does not guarantee they will win. It gives them a weapon.
If you are an established mineral or crystal dealer sitting on decades of knowledge, supplier relationships and inventory, refusing to learn how to use that weapon because you think AI is a fad would be an astonishing waste of the advantage you already have.
AI is not coming for the rock business by itself.
The entrepreneurs using it are.
And if you are in the gem and mineral business and refuse to adapt, there is a reasonable chance that somewhere right now, a younger competitor with less money, less inventory and far less experience is using these tools to learn how to compete with you considerably faster than you think.
In an industry built around fossils, becoming one probably should not be the business plan.