A shopper likes the jacket. An associate offers a virtual preview. Then comes a reasonable question: “What happens to my picture?” If the answer is a vague reassurance, the experience has introduced a new hesitation at precisely the moment it was meant to remove one.
For fashion retailers evaluating virtual try-on, privacy belongs in the commercial buying decision. A camera experience must be useful enough to start, understandable enough to trust and straightforward to leave. Our recommendation is to approve the shopper explanation and the underlying data handling together, before opening a pilot to customers.
This guide is for ecommerce, retail operations, CX and technology leaders evaluating iKawn Mirror or another fashion try-on service. It offers a practical review method, not a statement that every supplier has the same controls. Your privacy owner should determine the legal requirements for your markets, audience and actual deployment.
Hero image: an AI-generated editorial illustration of a shopper discussing an optional digital experience; not a photograph of a customer deployment.
Why this deserves budget before the launch date
In Cisco’s Data and Privacy Benchmark announcement, published January 26, 2026, 93% of surveyed organizations planned further investment in privacy and data governance. The study surveyed 5,200 IT, technology and security professionals with privacy responsibilities across 12 markets. Forty-six percent identified clear communication about data use as the most effective action for building customer confidence.
These are cross-industry professional opinions, not observed fashion shopper behavior or evidence of a conversion uplift. They nevertheless identify a procurement reality: AI projects compete for governance capacity as well as software budget. Our interpretation for a fashion buyer is practical. Reserve time for the person who must approve the data flow, the colleague writing the customer explanation and the store team that must answer questions. A subscription quote does not cover that internal work automatically.
Ask for these answers while the pilot scope is still adjustable. If a proposed save-and-share feature creates unresolved retention or access questions, a more limited experience may be a better first purchase. That is a scope decision to settle with the supplier, not a promise that removing a button removes stored data.
Separate the photo, the result and the profile
Two public examples show why “we use AI” says little about what happens to a shopper’s data.
Google Shopping’s try-on help page, checked October 5, 2026, says an uploaded photo is saved for reuse and can be deleted. It describes separate deletion routes for the original photo and generated looks. Google also says the photo is not used for training, while describing quality evaluation involving feedback and trained human reviewers with privacy precautions. This is Google’s account of that particular service, not a description of iKawn Mirror.
Zalando’s body-measurement feature announcement, dated July 20, 2023, said the photos used to predict measurements did not leave the customer’s phone and were deleted. This was a size-advice feature, not proof about every subsequent Zalando fitting-room experience. It illustrates a different design choice: the source image and the information derived from it need separate explanations.
The buying implication is to stop treating “the image” as one object. Ask about camera frames or uploaded photos, generated previews, any measurements or other derived information, usage records and contact details individually. Deleting one does not establish what happens to the others. Equally, a no-training statement does not answer whether operational storage or human access exists.
Ask for a data journey your team can explain
Request a one-page description of the proposed deployment, with the retailer, try-on supplier and any processing providers named. It should follow the shopper from starting the experience to leaving it. For each kind of data, record the purpose, processing location, people or services with access, retention rule and deletion route.
Make the supplier resolve these distinctions in writing:
- Processing versus storage: does input leave the device, even when a recording is not saved? Which service receives it?
- Immediate use versus later use: is anything retained for troubleshooting, quality review, analytics or model improvement? What can the retailer configure?
- Visible removal versus deletion: does ending the session only clear the screen, or also trigger deletion elsewhere? What happens to backups and existing exports?
- Anonymous-looking versus linked: can session records be associated with an account, device identifier, order or contact record? Avoid calling them anonymous without confirming that claim.
Ask for specific retention periods or defined deletion events, including any exceptions, rather than “only as long as needed.” Have your privacy and security owners review regional processing, responsibilities, supplier terms and the evidence supporting each claim. Do not infer these answers from whether the experience uses a live camera or a photo upload.
This extends the data-handling section of our virtual try-on vendor evaluation guide. Keep the resulting record attached to the actual pilot proposal so a later feature or provider change can be reviewed against it.
Make the shopper invitation match the approved service
For a first pilot, we recommend an optional try-on invitation with a clear route back to ordinary shopping. Before asking someone to start, explain what they will get, what input is needed, what happens to it and where they can read more or ask for help. Let the shopper see the value and the choice together.
Do not publish a universal promise such as “nothing is stored” unless it is supported for the whole agreed flow. Instead, write the explanation from verified facts: whether input is sent to a service, whether previews remain available, how long relevant records are kept and what ending the session does. Have store associates use the same explanation as the screen.
Keep the decisions to try on, save a result, share it and receive marketing distinguishable in the proposed journey. They serve different shopper purposes. Our commercial recommendation is to avoid making an email address the price of seeing the first preview unless that requirement has a defensible purpose and has been reviewed. Otherwise the pilot measures willingness to provide a lead as well as willingness to use virtual try-on.
The tradeoff is real. Saving a look can make a return visit easier; avoiding persistence can reduce later data-handling work. Asking for contact details may support follow-up; it can also complicate the participation decision. Choose the behavior that serves the pilot’s commercial objective, then verify the corresponding controls.
Rehearse the moments a polished demo can skip
Before customer access, run a rehearsal with consenting adult staff and test data. Use the proposed retail device or ecommerce journey, not a different showcase environment. Give CX or retail operations responsibility for recording the shopper-visible outcome, and ask the supplier to provide evidence for behavior that cannot be observed on the screen.
- Decline: refuse the camera or photo step. Confirm that the shopper can continue browsing or ask for ordinary assistance without repeated pressure.
- Interrupt: close the tab, lose the connection or leave the kiosk idle. Confirm the defined timeout and recovery behavior, including what remains visible.
- Change shopper: end a session and start the next one. Check the display, recent previews, back navigation and any download or sharing links for exposure of the previous participant’s content.
- Request deletion: use a test participant’s request to trace which records are removed, which remain under a documented exception, who responds and what evidence closes the request.
At a store, also walk around the installation. Check whether a person in the aisle could appear in the capture area or read the display. Plan placement and staff guidance accordingly. A successful screen reset cannot prove server deletion, and a written retention policy cannot prove that the next shopper sees a clean screen. The acceptance record needs both forms of evidence.
Judge participation without building an unnecessary photo archive
Choose the minimum measurement needed to answer the commercial question. A proposed pilot can count invitations, starts, successful previews and continuation to the chosen next action without making a permanent shopper-image archive its measurement plan. Have the team review the identifiers and retention behind those counts as well.
Record questions and failure categories without copying personal images into routine reports. If starts are low, interview willing participants about clarity, relevance and effort; do not assume privacy is the cause. If the experience is understood but declined, respect that outcome. Changing the wording to conceal a data practice would invalidate the trust you are trying to assess.
For approval, use a short evidence record: requirement, written supplier answer, observed result, accountable owner and unresolved issue. Our recommended launch blockers are an unexplained data destination, an unsupported customer promise, another shopper’s content remaining accessible, or no workable route for handling a data request. Assign a correction and retest before customer use. This is an editorial acceptance framework, not a regulatory certification.
Bring the real privacy questions to an iKawn Mirror demo
iKawn Mirror offers virtual try-on for fashion discovery across ecommerce, stores and brand activations. Use a demo to evaluate the proposed customer journey with your products, then request a deployment-specific account of processing providers, retention, access, optional saving and sharing, and deletion responsibilities. This article does not establish a universal Mirror data policy or certify a particular configuration.
Bring the intended channel, whether the device is shared, your shopper audience and the data rules your team must meet. Use the Mirror workflow guide to frame the experience, and involve your CX and privacy owners in the acceptance discussion. The useful outcome is an agreed experience your team can demonstrate, explain and support. Book a live Mirror demo to discuss that scope before committing to a customer pilot.