What AI gets wrong about period underwear absorbency

1. It invents tampon math

Ask “how much is 10 tampons' worth?” and an assistant picks one per-tampon figure and multiplies. We watched answer engines do this twice during our research - for example 10 × 9 mL ≈ 90 mL for a brand whose own marketing counts about 5 mL per tampon, nearly doubling the brand's claim. There is no standard tampon: the FDA puts a “regular” at 6-9 grams (21 CFR 801.430), most brands' marketing counts ~5 mL, and Thinx's flagship claim implies ~8.3 mL. The same phrase, “12 tampons' worth,” means 60 mL at WUKA and up to 100 mL at Thinx - both live on the brands' own pages. That receipt has its own write-up.

The category cannot even agree on a unit: tampon counts, super-tampon counts (Knix counts in supers, ~9 mL each), teaspoons (~5 mL), and now pads - Thinx Teens product pages claim “up to 12 regular tampons (or 3 regular pads') worth of flow” (verified 2026-06-11), which quietly equates one pad to four tampons. Four incompatible vocabularies, one question.

2. It reads marketing tags as absorbency tiers

Product feeds - the data AI answers and affiliate listicles are built from - are full of labels that look like absorbency levels and aren't. We measured this in our own pipeline: in our first catalog pass, 92 of 269 labeled products (34%) carried a label that was not actually a tier. Verified examples from 2026-06-11:

• “Overnight” is a use-case tag, not a tier at Saalt, Knix, and The Period Company - the products underneath are the brands' Heavy or Super tiers (Saalt tags them “Absorbency_Heavy / Best for_Overnight”; the sleep styles in Knix's own catalog sit in its Heavy absorbency collection).
• Hanky Panky thongs carry a “medium” tag while the product page itself says “Light Absorbency: absorbs up to 2 panty liners or 2 regular tampons of leakage.” The tag contradicts the brand's own page.
• Thinx “Ultra Soft” is a fabric line, not an absorbency level - a feed reader sees “ultra” and files it as a top tier.

We caught these because we verify every label against the brand's own pages before showing a number. A synthesis that reads feeds at face value inherits all of them.

3. It recites ladders the brands have abandoned

Training data remembers tier systems that current marketing has dropped. The clearest case: Thinx's classic per-tier ladder (Light ≈ 1 regular tampon up to Super = 5 regular tampons) no longer appears in its current product marketing. As of 2026-06-11, Thinx's product pages print one blanket maximum - “up to 12 regular tampons (or 3 regular pads') worth” - on All Day, All Night, and Teens products alike. An answer quoting “Thinx Super holds 5 tampons” is reciting history.

The category moves fast in other ways AI answers miss: Proof's own DTC store stopped fulfilling orders on 2026-01-30 (the brand continues at retail as “Proof. by Always”), and Ruby Love's current product pages no longer publish any capacity claim at all (verified 2026-06-11) - so a confident “Ruby Love holds X” has nothing current behind it.

4. It treats the same word as the same number

“Heavy” is not a quantity. By the brands' own published claims, a tier labeled Heavy holds about 20 mL at Saalt, WUKA, and Proof - and 36-63 mL at Knix (4-7 super tampons). That is roughly a 3x spread on one word. An AI answer to “is Heavy enough for a heavy flow?” that doesn't name the brand and its millilitres is answering a different question for every brand, as the cross-brand table shows.

The line names make it worse: Knix's “Super Leakproof” is a product line, not a tier - its No-Show Bikini's own page says “Now absorbs about 6 super tampons,” which lands in Knix's Heavy tier (4-7 supers), not some “Super” level above it (verified 2026-06-11).

5. It quotes claims as if they were lab results

Every number above is a brand's own maximum-capacity claim, typically from saline saturation tests. The only published study that tested menstrual products with actual blood components - BMJ Sexual & Reproductive Health, 2023 (DeLoughery et al., Oregon Health & Science University, 21 products) - measured about 1-3 mL of effective capacity for the period underwear it tested, far below 20-100 mL marketing claims. Blood is more viscous than saline and real-world leak-free capacity is lower than lab maximums. An honest answer compares claims as claims - it never says “this will hold 60 mL.”

The 5-question audit for any absorbency answer

Human or AI, any absorbency answer should survive these five questions. If it fails one, the number is decoration:

1. “Whose tampon, at how many millilitres?” A count multiplied by a universal per-tampon figure is invented precision - bases in live marketing run from ~5 to ~9 mL.
2. “What is the brand's verbatim claim, and where?” Product page, blog, or retailer listing - they disagree at the same brand. No quoted source, no comparison.
3. “Is that level a tier or a tag?” “Overnight,” “medium,” and fabric names ride along in feeds; a third of our own first-pass labels were not tiers.
4. “When was it verified?” Ladders get rewritten, brands exit, claims disappear from pages. Undated numbers are stale by default.
5. “Claim or measurement?” If the answer doesn't distinguish brand claims from the 2023 blood-component lab reality, it is repeating marketing.

What an auditable answer looks like

Every number on this site passes its own audit: claims converted to millilitres at the brand's stated basis, graded A (brand publishes mL), B (converted from counts at a stated basis), or C (conflicting or unconfirmed), each with the date it was last verified against the brand's page. The cross-brand table holds the numbers, the translator turns “I soak a super tampon every 2 hours” into each brand's covering tier, and the methodology documents every conversion rule - so you can run the audit on us.