Video · Coreplex
Why Supplement Shopping Lost Its Meaning — An Intro to Coreplex
How platform compliance flattened supplement discovery, why people stopped buying for each other, and how Coreplex uses governed, evidence-backed product records to bring meaningful discovery to agentic commerce.
Full transcript
Click any timestamp to play from that point.
Why we stopped recommending supplements
Okay, let's jump right into this. Have you ever noticed that we used to constantly recommend health products and supplements to the people we love and
now, well, we kind of just don't? Today on the explainer, we're unpacking a really fascinating shift in how we shop.
We are focusing specifically on why supplement shopping completely lost its meaning. It's a story about the internet, shifting rule books, and what
exactly happens when the human element of discovery just gets stripped away.
But more importantly, it's about how new technology might actually bring all of that back. To really wrap our heads around this, we're looking at insights
from a guy who spent over 15 years at iHerb. He actually started out as their very first webmaster. And look, his job back then was wild. He literally hand
typed product descriptions straight from the physical supplement labels, 20,000 of them. He read 20,000 different products, one at a time, and often way
more than once. I mean, just imagine the sheer mind-numbing absurdity of processing that volume of information every single day. But here is what's
really interesting. The human behavior that came out of all that reading, it didn't just make him better at his job.
It made him the absolute go-to guy in his personal life. Someone's knees hurt, he'd say glucosamine. Digestion off, probiotics. He was literally putting
together little care packages of CoQ10, red yeast rice, and fish oil for his uncle's heart health. And the crucial point here is that he wasn't alone in
doing this. Supplements are one of those rare categories where people regularly buy for somebody else. You care about your family, right? So if something might help your mom or your brother, you learn about it and you buy it for them.
When labels had meaning
When labels had meaning. Back in the golden age of internet discovery, people were actually learning about what they bought. See, in the old way, products
were talked about for what they actually were. You learned the origin of an ingredient, how it was traditionally used, the founder's backstory, maybe even
the real research behind it. There's actual legitimate research around red yeast rice and cholesterol, for example. And brands used to proudly point to it. That rich context created
curiosity, and curiosity sparked discovery. But today, the new way is totally stripped down. It's clinical, almost intentionally plain. You get a
product name, a list of ingredients, and just a wall of warnings and disclaimers.
All that rich ingredient history and traditional use completely gone. So, why did this happen? Let's move to section two, the great flattening. This is where
The great flattening
compliance got very, very serious across the internet. Now, to be fair, a lot of this needed to happen for consumer protection, but the way it rolled out
effectively flattened the entire internet. We didn't just regulate supplement commerce. We platformized compliance. Every single platform
created its own completely different rule book. The FDA has their rules, sure, but Amazon has different ones.
Google isn't Amazon, and Walmart definitely isn't Google. It basically turned e-commerce into a giant game of compliance bingo. What might be perfectly fine to say on your own
website could literally get your product pulled down somewhere else. So brands did what they had to do. They optimized everything. The label, the product page,
the digital shelf, strictly for the permission to sell. We all got incredibly good at answering just one single question. What are we allowed to
say? But here is the hidden cost of all that optimization. Let's walk through this chain reaction. Step one, nobody learns anything from a sterile, plain
label. Step two, because they aren't learning, nobody recommends anything to their loved ones anymore. Step three, fewer people buy for somebody else. And
step four, carts get smaller. There are just fewer things per order. By losing the storytelling and the context, the industry accidentally killed the exact
buying behavior that fueled the whole system. Protecting consumers is absolutely crucial, but it shouldn't require stripping every single product down until there's nothing meaningful
left. What should a machine be allowed to say? Better labeling and cleaner formulations are undeniably good things.
But the people making legitimate, high-quality products still need a responsible way to explain what they made, why they made it, and why you should care. As we move into a world
where AI and algorithms dictate what we see, how do we teach machines what actually counts as evidence versus what's just marketing fluff? That brings
Rebuilding agentic commerce
us to the technological cure. Section three, rebuilding agentic commerce. You see, the creator of our source material built a tool called FeedZip to clean up
supplement data for machines. It worked beautifully, but it led to a huge realization. Cleaning the data is actually the easy part. The hard part:
capturing meaning. So, let's define what we mean by agentic commerce. Think of it as a totally new generation of shopping powered by AI assistants, universal
commerce protocols, and machine-readable product infrastructure. It is fundamentally changing how products move across the internet. It gives us a real chance to rebuild that discovery phase
we lost. And honestly, this matters directly for your own shopping. If machines are going to help you discover, compare, and recommend products for your family, we can't just teach them the
price and whether it's in stock, we have to teach them what the product actually is, so you get recommendations that aren't just accurate, but meaningful.
This is exactly where a system called Coreplex comes into play. Now, a product can travel with a really rich structure rather than just a plain title and a low-
res image. It travels with identity: what exactly is it? Provenance: where did it come from? Evidence: what is the actual research behind it? And context:
why was it even formulated in the first place? With the right governance, the true full story of a product can travel right alongside it, directly into these
new AI shopping surfaces. But, okay, here is what makes this system responsible. And this is super important. Watch what happens when the
system doesn't know enough. In the past, an algorithm might just guess, right? Or marketers might just cleverly write around the missing facts. But to truly
protect consumers while still allowing for those rich stories, the system has to be rigorously, stubbornly honest about what it doesn't know. Let's look
at the live path of how this actually plays out. Step one, the system straight up declines to publish unverified claims. No exceptions. Step two, it specifically names the missing evidence.
Step three, it creates a visible gap in the product's record. We aren't hiding uncertainty here, and we certainly aren't writing around thin data.
Instead, in step four, the brand can see exactly what's missing. They are empowered to go in, fix it, and improve the record. That gap stays completely
visible until somebody puts in the work to fix it. Underneath all of this, the system stores knowledge as individual bite-sized atoms. Ingredient, form,
claim, evidence, and safety. Every single one of these atoms has its own independent review state. And there is one golden database rule that decides what the AI engine can actually touch.
It must be published and verified. This isn't just a polite suggestion or a loose style guide. It is strictly enforced deep down in the database
architecture. So the core philosophy is captured perfectly right here.
Uncertainty never quietly becomes certainty just because the system needs an answer. If a safety atom fails its review, it can disqualify a product
outright. Full stop. Uncertainty stays uncertainty. This kind of strict, unwavering governance is exactly what allows us to trust the information being
presented by AI assistants without having to strip away the product's soul.
Better stories with receipts
It's the ultimate consumer protection, but it still leaves the door wide open for real discovery. Which brings us to our final section today. Section four, better stories with receipts.
Ultimately, a product shouldn't enter the next generation of commerce completely stripped of the reason it exists in the first place. The colors, the origin story, the founder explaining
their passion, the ingredient history, the actual scientific evidence, all of that should travel with the product across whatever new platforms come next.
And this brings us to a crucial balance.
Governance makes sure that discovery actually deserves your trust. Think about it. Search fulfills a mission you already have, finding a specific item.
But discovery, discovery expands the mission. By using these governed atomic records, those rich product stories can safely travel across websites,
marketplaces, and AI agents. They aren't wild and uncontrolled. They aren't exaggerated. They are governed. We finally have a responsible way to tell
these stories again. Which leaves us with this final, really exciting thought.
If we can combine the rich historical context of products with the strict verified governance of agentic commerce, with better stories and actual receipts,
will people start buying for each other again? Will we finally get back to that beautiful human behavior of learning, discovering, and caring for our loved
ones through the things we recommend? It is a truly fascinating possibility for the future of shopping. Thanks so much for joining me on this explainer. Keep learning, keep questioning, and I'll see you next time.