Customer Service Automation for Beauty and Skincare Stores
In most retail the customer knows what they want and asks whether you have it. In beauty they ask what they should buy. “Which one for oily skin.” “Something for pigmentation.” “Is this safe while pregnant.” So the recommendation is the sale, and the claims you are allowed to make set the limit on what can be automated.
This page is for online beauty and skincare stores anywhere. It starts with the six workflows worth automating first, each linked to its own page with a description you can paste into AeroChat. Then it covers the messages that arrive, the channels customers use, the public data, two recipes you can copy in today, and the plan a store of your size typically fits.
Workflows you can automate
Each one links to a page with the flow, the description to paste, what to connect and what you still decide. The reasons for this order are on the ranked beauty store workflows page.
- Answer “where is my order” from your store. Checks the customer’s email or phone first.
- Recommend products by skin type and concern from your catalogue tags. Up to three, with one line on why each fits.
- Answer ingredient questions from the published list. Skin conditions and pregnancy questions go to a person.
- Handle returns and exchanges within your policy. Opened products and damaged items go to your team.
- Answer restock and subscription questions. From your store data and the restock dates you list.
- Hand reaction reports to a person straight away. Collects the product, the order and a photo first.
What automation does for a beauty store
- Recommendations at the moment of choice. A shopper comparing two serums at 11pm gets three suitable products and the reason for each, drawn from your catalogue.
- Consistent answers on ingredients and authenticity. The same factual answer every time, from the ingredient list and the sourcing statement you wrote once.
- Health questions handled safely. Anything about a skin condition, pregnancy or a reaction goes to a person, so the business never puts a medical claim in writing by accident.
The messages a beauty store gets
| Message | What the workflow does |
|---|---|
| “which one for oily skin” | Filters products by skin type, concern and texture, and recommends up to three |
| “is this fragrance free” | Answers from the published ingredient list |
| “what shade for medium skin, warm undertone” | Narrows to two or three shades, explains the difference, offers a person or a sample |
| “is this authentic” | Sends your sourcing and authorised-retailer statement |
| “when is my order coming” | Verifies email or phone, then reads the order and tracking |
| “safe for pregnancy?”, “will this clear my acne” | Gives only published product facts, declines to advise, hands to a person |
Recommendations come from your product data
A good beauty recommendation is a filter over attributes: skin type, concern, key ingredients, texture, budget, and whether it layers with what the customer already uses. If your products are tagged with those attributes, a conversation can narrow fifty products to three and explain why. If they are not, any system will answer from marketing copy, and marketing copy says every serum suits everyone. For most stores, the first job is a week spent tagging the catalogue. That work also improves your site search and filters.
The claims line
In most markets, a cosmetic cannot be advertised as treating or preventing a disease or medical condition. The EU Cosmetics Regulation, the US Food and Drug Administration and the ASEAN Cosmetic Directive all draw that line, and a product sold with a treatment claim can be treated as a medicine. A chat reply is written, timestamped and attributable to your business, so an automated answer to “will this fade my melasma” is a claim you are responsible for.
The correct behaviour is narrow. State the facts you publish, such as the ingredient list and the manufacturer’s directions. Decline to advise on the condition. Pass the message to a person, and outside hours say when a person will reply. In AeroChat you set this as a handover rule, and the system applies it however many times the customer asks. See live chat handover.
Shade matching
Shade is a common reason a beauty order comes back, and it cannot be settled in text. The useful automation narrows: ask about undertone and depth, offer the two or three shades those answers point to, explain the difference, and offer a person or a sample for the final call. It should never pick a shade from a photo and assert it. Lighting makes that unreliable.
Channels customers use
Instagram brings discovery and recommendation conversations, often from a post or a story reply. AeroChat answers both direct messages and post comments there. WhatsApp carries repeat customers and order questions. Web chat catches the shopper comparing two products on your site. AeroChat’s TikTok channel covers TikTok direct messages and TikTok Shop buyer chat in the same inbox. All of it should land in one place, because a customer’s history makes the next recommendation better. A customer who returned a shade last month should not be offered it again. See shared inbox.
What the data says
| Figure | What it measured |
|---|---|
| About US$450 billion | size of the global beauty industry |
| 7% a year | growth from 2022 to 2024 |
| About 5% a year | expected growth through 2030 |
Source: McKinsey & Company, State of Beauty 2025, 2025.
| Figure | What it measured |
|---|---|
| 67.7% | of consumers agree a response from an AI chatbot is helpful |
| 72.4% | are more likely to buy from a brand that offers messaging |
| 79.3% | want proof a business is legitimate before engaging with it |
Source: Meta, State of Business Messaging. Kantar survey of 11,056 adults in 22 markets, commissioned by Meta, which owns WhatsApp, 2025.
Slower growth means more competition for the same shopper. The store that answers “which one for me” well, and answers “is this authentic” plainly, keeps the sale. The 79.3% figure is why the authenticity answer deserves care.
Recipe: product recommendations with the claims rule
Paste this into AeroChat (Skills, then Build from a description) and change the example values to yours.
Copy this into AeroChat
When a customer asks which product suits them, follow these steps.
- Ask their skin type, their main concern and their budget if they have not said.
- Recommend up to three products from our catalogue whose tags match, and give one line on why each fits, based only on the product’s listed ingredients and description.
- If they ask about an ingredient, answer from the product’s ingredient list.
- If they ask whether a product is authentic, send our sourcing statement from the knowledge base.
If they ask about a skin condition, a medical treatment, pregnancy, breastfeeding, an allergy or a reaction, do not advise. Share only the ingredient list and the manufacturer’s directions, say that our team will reply, and hand the conversation to a person. Never say a product treats, cures, clears or heals anything. Reply in the customer’s language.
What to connect:
- Your store. Shopify is built in: install the AeroChat app, which syncs products and tags. WordPress stores connect the same way.
- Your sourcing statement. Uploaded to the AeroChat knowledge base.
- Your channels. Instagram, WhatsApp and web chat. Built in.
Recipe: shade questions
Copy this into AeroChat
When a customer asks which shade to choose, ask whether their skin is light, medium, tan or deep, and whether their undertone is cool, warm or neutral. Offer the two or three shades in that product’s range that match, using our shade guide, and explain the difference between them in one line each. Offer our sample sachet set for US$5 (example price), or offer to pass the conversation to our team for a final recommendation. Never choose a shade from a photo.
What to connect: your store, and your shade guide uploaded to the knowledge base. Nothing else is needed.
After you paste a recipe, AeroChat builds it as a draft with a checklist: connect your accounts, upload your lists, then switch it on. Any keys are entered in the dashboard and never pass through the conversation.
Using a different system? If it has a public API, it connects to AeroChat the same way: describe what you want, paste in its key, and it runs.
Can’t find your tool? Message us in the chat on this page and we’ll tell you how it connects.
What a store of your size typically invests
AeroChat plans are priced on AI responses, the individual replies the AI sends. A recommendation conversation usually takes several responses. A worked example: a store with about 15 recommendation and order conversations a day, at around four replies each, uses roughly 1,800 responses a month.
- A new or small store usually fits Starter: 500 AI responses a month and 3 team seats.
- A store at the volume in the example usually fits Growth: 2,000 AI responses a month and 10 seats, plus payment links in the chat, WhatsApp broadcasts and insights.
- A larger store or a brand selling across several markets usually fits Advanced: 5,000 AI responses a month, 25 seats and auto-translation.
If a launch or sale month uses more, you top up with extra AI credit. What the investment does is answer “which one for me” while the shopper is deciding, and keep health questions with your team.
Test it against your own catalogue
Install AeroChat, paste the recommendation recipe, and ask it what you would recommend for oily, acne-prone skin. Then ask whether one product is safe in pregnancy and check it hands the question to you.
Related reading: live chat handover, AI hallucination, WhatsApp message templates, shared inbox, how to sell beauty products online. All beauty store workflows. All industries.
Install AeroChat from the Shopify App Store and start your 7-day free trial.
Not on Shopify? Sign up here.