Customer says
“I used the retinol serum twice and woke up with a rash all over my face.”
Cosmetics · supplements · personal care
Tamper, contamination, adverse reaction — the AI knows when to triage and when to disengage and route to safety.
Cosmetics and supplements have one rule: never debate a safety claim. The AI handles tamper, contamination, color-mismatch, and shade-issue claims with proper photo flows — and the instant a customer mentions a reaction, it hands off to your safety queue with no further engagement.
Tamper detectionContamination routingAdverse-reaction safety queueShade mismatchAge-verification gate
Free plan included · 5-minute install · no credit card
The problem
Why a generic AI is dangerous for cosmetics + supplements
A customer says 'my face is breaking out from this serum.' That's not a returns conversation — it's a regulatory event. Generic returns AI tries to triage: ask for photos, suggest patch testing, debate ingredients. All of that is wrong, all of it creates legal exposure. Eturns recognizes adverse-reaction language at the first message, hands off to your safety team within 60 seconds, stops engaging, and never makes a medical or causation claim. The same AI handles cosmetic-only claims (tamper, contamination, shade) deftly because the regulatory bright line is built in.
How a real claim flows
One conversation, three roles played: customer, AI, your team.
AI handles
Recognizes 'rash' as adverse-reaction language. Sends a brief acknowledgment ('I'm sorry to hear this — I'm connecting you with our team right now'), hands off to legal + safety queue with the order context attached, stops engaging.
Outcome
Customer reaches a human within minutes. Legal team handles the conversation, evaluates the safety significance, and decides on next steps with appropriate documentation. AI never debated ingredients or offered medical guidance — exactly the right behavior.
60sfrom adverse-reaction language to human safety-team hand-off
What the AI knows
6 defect categories tuned for cosmetics.
Each category controls which photo the AI requests, which department the case routes to, and which SLA the customer hears. Turn any of them off in your merchant dashboard.
Adverse reaction (immediate safety-queue hand-off)
adverse_reactionCustomer reports irritation, breakout, allergic reaction, supplement side effect. AI hands off to your safety queue within 60 seconds. Never debates ingredients. Never asks for photos. Never offers medical advice.
Tamper / seal broken on receipt
tamper_detectedOuter seal, inner safety wrap, or shrink band broken before customer use. AI requests packaging photos and routes to fulfillment + safety joint review.
Visible contamination
contaminationProduct separated, foreign material, mold, off-color. AI routes to safety queue with photos for retain-and-investigate before any reship.
Shade / color mismatch
color_mismatch_to_listingFoundation, lipstick, hair color out of expected range. AI asks for swatch photo against listing image, processes as a standard return / exchange.
Expired or near-expiry product
expiry_violationPast best-before, or expires within merchant's freshness window. Auto-approves replacement.
Age-verification failure
age_verification_failureFor age-gated SKUs (retinol > 1%, certain supplements, CBD where regulated). AI handles the verification check at the start of any return on these SKUs.
Where it routes
Specialist destinations, not a generic CS queue.
Eturns ships with 14 routing destinations beyond CS. The cosmetics disclosure profile uses these:
| Department | Triggered when | SLA the customer hears |
|---|---|---|
legal | Adverse reaction, contamination, undisclosed-ingredient claim | Hand-off to legal + safety team within 1 hour |
warranty | Tamper, expiry violation, packaging defect | Replacement approved within 4 business hours |
fulfillment | Shade / color mismatch — exchange path | Same-day exchange shipment |
Disclosure profile
Standard with mandatory safety overlay
Adverse-reaction language is a hard tripwire: AI hands off to the safety queue and stops engaging. No 'have you tried patch testing.' No 'maybe try a smaller amount.' No 'let me check the ingredients for you.' The legal posture is the right one — there's no upside to AI commentary on a possible medical event, and there's significant downside.
Who this is for
If your store sells any of these, the cosmetics profile fits you.
- Cosmetics DTC brands (skincare, color, hair)
- Supplements + wellness DTC brands
- Personal care + intimate-wellness brands
- CBD / hemp / age-gated retailers
Questions cosmetics merchants actually ask
Frequently asked.
Does the AI ever make ingredient or causation claims?
Never. On any safety-class language, the AI hands off and stops engaging. There's no scenario where an AI commenting on whether a specific ingredient caused a specific reaction is the right answer.
What about minor cosmetic concerns (color, scent, texture)?
Those are normal returns. AI handles them with photo evidence and the standard exchange-or-credit flow. The bright line is anything implying physical reaction or contamination — that's an instant route, not a triage.
How does age verification interact with returns?
On age-gated SKUs, the AI runs the verification check at the start of the return conversation. If verification fails, the return is held for manual merchant review. Adult-21+ supplements and CBD merchants need this gate to stay compliant.
Ready to give cosmetics the returns flow it deserves?
Install Eturns from the Shopify App Store. Free plan included. Setup takes 5 minutes — the AI auto-detects cosmetics from your products and policies.