Clair supplies the headline. This page tests what backs it. Claim by claim, the public evidence falls short of the hormone-monitoring pitch.
8 claims shown
01Company-reported result
94.1% accuracy
WHAT IS CLAIMED
The archived FAQ reports 94.1% cycle-phase classification accuracy.
WHERE THE PITCH BREAKS
94.1% is a classification score. It does not report hormone-concentration error. Paired assays and subject-level cross-validation do not change what the score measures. Selling the percentage as proof of a quantitative hormone readout skips the test that matters.
WHAT CLAIR MUST SHOW
Publish the full methods, confusion matrix, confidence intervals and hormone-concentration errors.
Clair describes a multi-modal system using more than 130 biomarkers.
WHERE THE PITCH BREAKS
A longer input list is not stronger validation. Clair must show that its model separates hormones from illness, exercise, sleep and other influences. Counting features does none of that work.
WHAT CLAIR MUST SHOW
Identify which signals improve performance and publish where the model fails in real life.
The archive describes a prototype cohort, then a plan to collect and validate data using production hardware.
WHERE THE PITCH BREAKS
Clair changed the device. The new device needs its own validation. Its archived study plan puts that validation in the future. The prototype’s score is not a transferable certificate for the bracelet customers will receive.
WHAT CLAIR MUST SHOW
Publish independently evaluated results from the finished bracelet.
Clair’s research poster names mcPHASES and lists Fitbit, Dexcom and Mira equipment in its methods.
WHERE THE PITCH BREAKS
The poster’s methods do not include a Clair bracelet. Modeling someone else’s recordings does not test Clair’s hardware. Put the actual product in the experiment.
WHAT CLAIR MUST SHOW
Test the production bracelet against reference hormone measurements and publish its errors.
An app screen shows estrogen at 268 pg/mL beside a claim of validation against gold-standard hormone assays in a real-world clinical trial.
WHERE THE PITCH BREAKS
“Validated” is printed on the slide. The trial identifier and hormone-specific errors are not. A phase score cannot certify the concentration on an app screen. This redacted public deck leaves that claim uncheckable; it does not reveal what investors received privately.
WHAT CLAIR MUST SHOW
Name the trial, tested hardware and participants. Publish the concentration errors behind the displayed readouts.
Mayo independently lists Clair in its Accelerate multiyear program.
WHERE THE PITCH BREAKS
The Mayo relationship is real. Using it as proof that the bracelet works is the leap. An accelerator announcement supplies no finished-device validation. The logo cannot run the experiment.
WHAT CLAIR MUST SHOW
Specify the data, the task and the validation results produced by the collaboration.
a16z speedrun advertises $18 million in signed B2B contracts. Clair’s own public fundraising deck describes device-purchase MOUs and partner LOIs. Reach Capital separately describes 100+ clinic LOIs.
WHERE THE PITCH BREAKS
The $18 million headline is easy to repeat. The obligations behind it are not available in this deck: MOU counts, values and partner names are redacted. “MOU” and “LOI” alone settle neither enforceability nor revenue. A headline is no substitute for the terms.
WHAT CLAIR MUST SHOW
Identify the agreements behind $18 million. Disclose purchase obligations, cancellation rights and cash collected.
Clair’s terms describe a wellness product and rule out diagnostic use, contraception, and use as the sole fertility-planning method.
WHERE THE PITCH BREAKS
The sales pitch does not cancel the terms. Clair excludes diagnosis, contraception and sole reliance for fertility planning. Put those limits next to the promise. Customers should not have to excavate them.
WHAT CLAIR MUST SHOW
List the supported uses, their clinical evidence and any required authorization.