A privacy-first virtual try-on lets a shopper see a piece on their own face without the store, or the software, keeping that face. The shopper agrees first, the photo is used for that one preview, and the photo and the result are gone when they finish. If a try-on tool keeps an archive of customer faces, it is not privacy-first, whatever its policy page says.
Jewellery is tried on in public: at a counter, often with family around, on a screen the next customer will use. That makes privacy a sales question as much as a legal one. A shopper who trusts the screen tries on more pieces.
What does privacy-first virtual try-on mean?
Four things, and you can check each one: consent before every photo; no stored photos, neither the one the shopper takes nor the result; shared screens that clear themselves; and no shopper faces in your marketing. Anything less leaves a pile of personal photos that someone has to protect.
Six checks to run on any try-on vendor
1. Is consent asked of every shopper, every visit?
A tick box at setup is not consent from the shopper standing at your counter. Look for a consent screen before the first photo of each visit, separate from other terms, with a clear way to say no and keep browsing. Ask whether the server checks consent before it creates anything, or only the screen does.
2. Are the photo and the result both deleted?
Some tools delete the upload but keep the generated image, which shows the same face. Ask where each one goes, how long it stays, and whether it ends up in backups.
3. Does a shared screen clear itself?
A kiosk that remembers the last shopper’s photo shows it to the next one. The screen should clear when a shopper finishes, and on its own after a few minutes without use.
4. Are photos used to train AI?
Ask in writing whether shopper photos are used to train or improve any model, the vendor’s or its AI provider’s.
5. What does the vendor keep, and for how long?
Some record is useful: proof that a shopper agreed. It should hold no photo, and it should be deleted on a schedule the vendor tells you.
6. Where do the marketing images come from?
Campaign images should be made with AI models or with people who agreed to it, never from shoppers’ try-on photos.
Red flags
A “save to gallery” button for shopper try-ons, consent buried inside general terms, a vague line about retaining images “to improve the service”, and a kiosk that shows the previous shopper’s face. Each one means customer photos are being kept somewhere.
How TryGehna answers these checks
Each shopper consents once per visit, before the first photo, and can say no and still browse the full catalogue. Our server checks that consent before it creates anything, and a consent lasts for that visit only, at most 12 hours.
The photo is used only to create the preview. Neither the photo nor the result is saved, in any gallery, on our servers or in backups, and neither is used to train AI. On a kiosk, everything clears when the shopper finishes, or after 5 minutes without activity. A shopper can download their own result; that copy is theirs.
When you send a piece to a customer’s phone, they can try it on once from the link under the same rules: consent first, nothing kept.
We keep one record: that consent was given, with the time, the policy version and the type of browser. Never the photo. Consent records are deleted after 90 days. Studio photoshoots are made with AI models, so no shopper’s try-on photo is ever used for your campaigns. The details are in Customer Photos and Consent in Virtual Try-On and on our privacy and terms page.
Want to see the consent screen and the reset for yourself? Book a demo or see pricing.