"Organic", "Made in Italy", "Doctor-Founded". Which of Your Claims Can a Machine Trust?
Organic, made in Italy, doctor-founded. Your claims sell to people. A machine needs to know who said each one and what’s published behind it.

In this article: the part of Interpretability Engine Optimization (IEO) that handles claims. Not whether a claim is true, which is your business and sometimes the law's, but whether a machine can tell who made it and what's published behind it. Its job is to attach to every claim a label that says where it came from, so the claim travels with its source instead of alone.
Here is what happens without it.
A shopper asks an AI assistant for "an organic cotton baby bodysuit." Your product page says organic cotton in the title. Your competitor's says GOTS-certified organic cotton, with the certificate number. To a shopper reading both, you look the same. To a machine deciding which one to recommend, one claim arrives with its evidence attached and the other arrives alone.
Now flip it. Another shopper asks for "a skincare brand founded by a doctor." Your About page says exactly that: founded by a dermatologist. But on the product, the assistant finds nothing, and it recommends a brand that wrote "dermatologist-founded" into its product data. The machine couldn't see that your claim was there at all. It saw one word and zero words.
Both failures have the same cause: nobody handed the machine each claim with its source.
Why claims are different from facts
"Cotton" is a fact about the product. "Organic cotton" is a claim: it asserts something a shopper can't check by touching the fabric. "Made in Italy" and "dermatologist-tested" are the same kind. They sell on trust, and they're exactly the kind of thing a machine should carry together with who said it.
Regulators already treat these words as weighty. In the U.S., most businesses that sell agricultural products as organic have to be certified under the USDA's organic program, and only certified ones can use the USDA organic seal. An unqualified "Made in USA" on a label falls under a Federal Trade Commission rule: the product has to be all or virtually all made in the United States. The rules exist because the words carry weight on their own.
Shopping assistants are now the ones repeating the words, to people who may never see your About page. If they repeat a claim as a bare fact, with nothing published behind it, the complaint lands on you. If they skip a claim you did publish, the sale goes to someone else.
The technical name: evidence and declared states
In IEO, every statement about a product or a brand carries a declared state, a label that says what it rests on. The Methodology names five:
- Explicit: the source says it in so many words.
- Derived: it follows from the source by a stated rule. "Crafted in our Prato workshop" → Made in: Italy.
- Inferred, with confidence: probable, with the basis and the confidence stated, and never promoted to fact.
- Unresolved: the source doesn't say. The field stays empty on purpose.
- Conflicting: two sources disagree, and the business hasn't decided yet.
Declared states make the source and the basis of each claim explicit. With them, your "organic cotton" is explicit: stated by the brand on the product page. If you've published a certificate, it travels alongside as supporting evidence. Neither version is a judgment on whether the cotton is organic. Both say exactly what you published and what you published to back it, and a machine can carry that forward as "the brand says" instead of as a fact of nature.
Three moves on the bodysuit:
- List every claim. Not every fact: the ones a shopper takes on trust.
- For each one, note where you say it and what else you've published that supports it. A certificate, a standard, a lab, a founder's bio, a factory address. Or nothing beyond the statement itself.
- Label it. Stated by the brand. Stated, with evidence attached. Derived from something you said. Never "probably".
Where it lands: each claim with its state
Here are four claims from the bodysuit brand's own product page, before and after:
| The claim | What the page says | State before | State after |
|---|---|---|---|
| Organic cotton | Title: "organic cotton"; nothing else published about it | A word | Explicit · stated by the brand on the product page · no certificate published yet |
| OEKO-TEX certified | Bullet: "OEKO-TEX Standard 100 certified", with the certificate number and a link to the certificate | A word and a link | Explicit · stated on the product page · supporting evidence: certificate number and issuer, linked |
| Made in Portugal | Bullet: "sewn in our Porto workshop" | A phrase about a workshop | Derived · "sewn in" names the making, Porto names the place → Made in: Portugal |
| Hypoallergenic | Bullet: "hypoallergenic", nothing else anywhere | A word | Explicit · stated by the brand · no test or standard published alongside |
Nothing was added and nothing was rejected. Four claims got labeled with where they come from. The first and fourth rows aren't failures. They tell you exactly what a shopper, or a machine, will find behind the word: your statement, and nothing else yet. Whether to publish more is your call, and now you know where that call is.
What to do this week
- Make a list of every claim on your best-selling product and your About page. Words a shopper trusts without checking: organic, natural, certified, tested, made in, founded by, award-winning.
- Next to each, write where you say it and what else supports it. A certificate, a standard, a lab, a bio, a factory. If there's nothing beyond the statement, write that.
- Attach the claim and its support to the product, not just to the About page: a certification field, a link, a standard's name, or the plain statement with its source.
- For claims with nothing behind them, decide whether to publish what supports them. Either way, the claim travels with its provenance: "stated by the brand." There's no separate place to keep a claim "as marketing"; the copy a person reads is copy a machine can read too.
- Watch for derived claims. "Woven at our mill in Donegal" supports "Made in: Ireland" because it names the making and the place. "Donegal" alone names a place. "European design" supports nothing about where it's made.
Practical example
A tea company's product page for a green tea says: "Hand-picked in Shizuoka. Organic. Rich in antioxidants. Loved by our customers for 12 years."
A shopper asks an assistant: "Is this tea organic and where is it from?"
Before you read on, give each claim its state.
Hand-picked in Shizuoka → Explicit for origin: Shizuoka. Japan → Derived, geographically. "Hand-picked" stays as a stated claim about process.
Organic → Explicit, stated by the brand on the product page. No certificate is published there, so that's exactly what travels: the statement and its source. The assistant can say "the brand describes it as organic." It shouldn't say more, and it doesn't need to say less.
Rich in antioxidants → Explicit, stated by the brand, with no measurement or standard alongside. Same treatment: carried as the brand's statement.
Loved by our customers for 12 years → Explicit, as stated. Don't turn it into the company's age or a review score; the page didn't say either.
That's the verified, never invented principle from section 4.2 of the IEO Methodology, with its declared states, applied to the claims a brand makes about itself: every one carried with its source, none invented, none rejected.
Check what's explicit, what's missing and what contradicts itself on one of your products → Run the free audit
To go deeper → The IEO Methodology, section 4.2 (doi.org/10.5281/zenodo.22416784) · nsolvia.com/ieo
Sources: U.S. Federal Trade Commission, Made in USA Labeling Rule, 16 CFR Part 323 · USDA Agricultural Marketing Service, "Do I Need to Be Certified Organic?" (National Organic Program fact sheet).
← Meta Built Your Ad. Your Best Selling Point Stayed in a Paragraph. · ⌂ IEO · Next: Free Delivery. But Will They Carry It Upstairs? →
Series: Interpretability Engine Optimization · Knowledge Domain: Interpretability (IEO)
Interpretability Engine Optimization (IEO)
The UNDERSTAND Layer
nsolvia.com/ieo