How does an Instagram boutique / online clothing seller train intusell's AI? A step-by-step guide

An Instagram boutique's DM inbox is always full of the same questions: "Which size is this?", "I'm 1.65 m, 60 kg — would it fit me?", "Is it in stock?", "Price?", "How many days for shipping?", "Are there color options?". You post a story and within minutes dozens of "price?" messages land. The post fills up with comments. You go live in the evening and hundreds of messages pile up in the "is it here yet, is it here yet" rush, and while you reply to one person, three customers say "they're not responding" and buy from another boutique. The customer who at 11 p.m. says "I want this right now, can you send the IBAN?" gets left until morning, and the impulse fades.

intusell takes over this load. But to do that, it first needs to learn your boutique, your products, and your sales style. This article explains how an Instagram boutique / online clothing seller trains intusell's AI from scratch, step by step. This is the Instagram boutique chapter of our sector-based "how to train your AI" series, and it's the main article of the series.

Quick answer

Why is training important?

intusell is not an off-the-shelf chatbot; it's a fully autonomous AI sales assistant that behaves like a senior DM sales advisor. A good boutique salesperson doesn't start on day one without knowing your collection either: they need to know which products you sell, your size and fit scheme, your shipping and return conditions, and which question they should leave to you. Training makes the AI do two things correctly: give the right information (size, price, fabric, shipping) and give it in the right way (your brand voice, your tone, your sales discipline).

An untrained assistant either gives very generic answers or guesses on something it doesn't know. In boutique sales, guessing is expensive: a wrong size recommendation means a return, made-up stock information means a "I placed an order but it was out of stock" complaint, and a wrong price loses the customer. A well-trained assistant, on the other hand, stops when it doesn't know, says so honestly, and connects to the live team. That's why the real aim of training isn't to make the AI talk a lot; it's to make it talk correctly and stay silent in the right place.

Who is it for?

This guide is for Instagram boutiques and online clothing sellers whose DM and comment inboxes are dense with size and stock questions, answering the same questions over and over:

• Women's/men's/children's clothing, shoe, bag, jewelry, and accessory boutiques selling only from Instagram or from Instagram + WhatsApp
• Sellers crushed under the DM load of live and story sales, working alone or with a small team
• Brands that can't keep up with the "price?" questions under comments and the incoming DMs
• Ready-to-wear, second-hand / vintage, designer boutiques, and online clothing stores doing dropshipping
• Those who want to re-engage customers who said "let me think about it," "let me ask my spouse," and disappeared

You don't need a technical team. All of the steps below are completed from sections in the panel, without writing code.

The structure of training: 5 core steps + 2 continuous steps

Splitting the training into two groups makes your job easier. The first 5 core steps get the assistant working; these steps are the one-time part of the setup. The following 2 continuous steps (label review and opening channels) are the ongoing part you sharpen as you use the assistant. You'll find all of them in order below.

1. Upload the knowledge base

The foundation of training is the knowledge base (RAG). In the panel, you upload all of your boutique's textual information to the Knowledge Base section. The AI automatically searches these sources whenever it produces a response and only uses the information written here. This is the technical answer to your "don't let the AI make things up" concern: the AI talks about whatever you hand it.

Supported source types:

Each file you upload is automatically chunked and made searchable with pgvector. When you upload a size/measurement chart CSV, the AI finds the right row and responds when a customer asks "which size for 1.65 m height, 60 kg?". When you upload a product list, it answers "do you have it in brown, is there any S left?" with the information in that list.

Must-uploads for an Instagram boutique: product names and descriptions, size/measurement chart (bust-waist-hip, shoe-size equivalents), fabric and care information, color and stock status, shipping times and fees, return/exchange conditions, payment methods (bank transfer/EFT, cash on delivery, installments), and the most frequently asked questions. When you upload these topics, the bulk of your DM inbox already becomes auto-answered.

An important note: if your product list changes often (new products, sold-out sizes, price updates), keeping the knowledge base current is in your hands. When you upload the up-to-date file, the AI uses the new information in its next response. The AI doesn't present "there's S size" from an old list as "current"; it knows whatever you gave it, and it says when it doesn't know. The more disciplined you keep your stock tracking, the higher the assistant's accuracy.

2. Set the persona and tone

In the first step you taught the AI what to say; now you'll determine how it'll say it. There are two settings in the panel:

• Assistant name (ai_persona_name): The name it'll introduce itself to the customer with. Most boutiques use a real advisor's name from the team or a warm name that fits the brand.
• Tone (ai_tone): Whether it'll speak in a friendly and casual way, an elegant and calm way, or a fast and energetic way.

A young streetwear boutique and an elegant evening-gown boutique won't have the same tone. The first can speak in a relaxed, witty, informal voice; the second in a polite, elegant, formal voice. Given the nature of Instagram, most boutiques prefer a friendly, informal voice — this setting depends entirely on your brand personality. The tone you set is reflected in the AI's message voice on every channel; the tone in Instagram DM stays consistent with the tone on WhatsApp.

Multilingual note: Whatever language the customer writes in, the AI replies in that language. When a customer from abroad writes in English or German, you don't do anything separately; the assistant detects the language automatically and replies in it while keeping your tone. This is a quiet advantage for boutiques that ship abroad.

3. Define the response rules

The persona defines "who it is," and the response rules define "how it'll behave." This is where you transfer your sales discipline to the AI. With the AI Manager Chat, you add rules by writing in plain language, just like instructing a teammate.

Typical rules for Instagram boutiques:

• "If you're not sure about size, stock, or price information, don't make it up; say you don't know and hand the chat to the team."
• "If the customer gives measurements, suggest the suitable size from the size chart, but don't say 'it'll definitely fit'; say 'these measurements usually fit size X.'"
• "When return/exchange is asked, read the policy clearly from the knowledge base, don't add commentary."
• "Once an order is finalized, communicate the payment methods and shipping time clearly; share the IBAN/payment information only the way it's defined."
• "If the customer says 'I'll think about it,' don't pressure them; remind them of the product's standout feature and leave it warm."

The most common mistake in boutique sales is the AI throwing out a size or stock it doesn't know "just to say something," and this turning into a return or a "I placed an order but it was out of stock" complaint. In intusell this boundary is two-layered. The first and main control is the system prompt given to the AI: from the outset, the AI is constrained so it won't go outside the knowledge base and invent size/stock/price. The second, on top of that, is a guardrail layer that scans responses and flags risky statements. This layer works in "shadow" mode by default: it detects and flags, but doesn't hard-block the response. So the system's reliability comes not from "a wall that blocks everything" but from the AI being correctly constrained from the start.

If you want to compare two different approaches, you can use the A/B test feature: for example, you put a softer suggestion and a clearer closing sentence side by side with 50%-50% traffic and measure which one converts to more orders.

4. Teach it from past DM and sales conversations

This is the step that moves training from "good" to "specific to your boutique." In the panel you upload voice recordings of your past sales conversations (MP3, MP4, WAV, M4A) and mark each one as Won or Lost — for example, did the customer complete the order, or did they leave saying "let me think about it."

The system uses these recordings in two ways:

That way the AI learns how your best salesperson explained fabric and fit quality to someone who said "it's a bit pricey," and how they kept warm and moved someone who said "let me ask my spouse" to an order. The KVKK side is protected: PII (personal data) masking and explicit consent are applied to uploaded recordings.

This step isn't mandatory, but don't skip it. The knowledge base teaches the AI "what" it knows; past conversations teach it "how you sell." When the two combine, the assistant truly resembles an experienced DM sales advisor.

5. Refine with label review (continuous step)

The first four steps get the AI working. From there on is the continuous part that perfects it over time. Every AI response lands in a label review queue. Here you can do three things: approve, reject, or correct.

The AI learns from these corrections. Say it described a size incompletely or referred to a product by the wrong name; you correct it, and in similar situations it now uses the right wording. Over time, patterns specific to your boutique build up (the product names you use, your standard campaign sentences, your brand voice like "I styled it," your return-routing style).

We recommend looking at this queue for 10-15 minutes a day during the first two weeks. Over this period the correction rate drops quickly, because the AI learns the frequently made mistakes. Review is the "live" part of training: the system gets smarter as it's used.

6. Open the channels (continuous step)

When the training is ready, you put the assistant in front of customers. intusell brings all channels together in a single inbox: WhatsApp, Instagram DM, Instagram comments, Facebook Messenger, Telegram, Web Chat, and email. Whatever channel the customer writes from, the same trained assistant responds with the same information.

Channel connection methods:

• Meta channels (Instagram, Messenger): connect with one-click OAuth.
• WhatsApp: connects in about 1 minute by scanning a QR code from your phone — no Meta Business approval required. The official Cloud API option is also available.

For an Instagram boutique, the right starting point is WhatsApp and Instagram DM; these are approved and usable right away. Most sales already flow into DMs with a "message us in DM" call; the assistant handles this traffic within seconds. Instagram comment automation ("price?", "is it in stock?" comments under product posts) is rolled out gradually depending on Meta approval. The channel gates are separate: the Instagram DM gate (instagram_dm_enabled) and the Instagram comment gate (instagram_comments_enabled) are independent; you can automate only DMs first and open comments when approval arrives.

There are operating modes for handoff to a human: ai_only (the AI answers everything), human_only (everything goes to the live team), hybrid (the AI normally answers and escalates to the team when needed). Most boutiques start with hybrid; the AI handles the repeated size, stock, and price traffic, and special requests or sensitive situations come to you. It's worth adding a note: a product photo, screenshot, or voice message the customer sends is not on its own a reason for handoff to the team; the AI tries to make sense of the attachment (for example, it tries to interpret an image that comes with "do you have this model?"), but hands off when it can't complete the task.

After you open the channels, the CRM and proactive follow-up kick in: the assistant can make an opportunistic re-offer to suitable WhatsApp customers being followed up who said "let me think about it" and disappeared, or who showed interest in a product whose size was out at the time (for example, when the size comes back or when a similar product appears in the new season). This is the part of boutique sales that leaves the most money but is the most neglected. The full details of day-to-day operation — from incoming DM to sale, follow-up, and handoff to the team — are in a separate article: how an Instagram boutique uses intusell.

How long does training take?

A working setup takes half a day:

1. Uploading the product list, size chart, and first files to the knowledge base: 1-2 hours (shorter if your existing Excel list is ready)
2. Persona, tone, and first response rules: 30 minutes
3. Connecting WhatsApp and Instagram DM: 1-5 minutes per channel
4. First test chats: 30 minutes

But there's no moment where "training is done." Over the first two weeks, as you approve and correct responses in the label review queue, the assistant sharpens to your boutique. Uploading past conversation recordings and running A/B tests are also ongoing improvements. The setup is fast; mastery is continuous.

What it isn't

Putting intusell in the right category matters, because a wrong expectation leads to a wrong setup.

• It is not a chatbot that gives canned answers. It's not a decision tree, but an autonomous sales assistant that represents your boutique to the extent you train it.
• It is not a tool that invents size/stock/price. It only uses the information you upload; it says when it's not sure and hands off to the team. This is the most important trust message.
• It is not an e-commerce infrastructure or payment system. It doesn't replace your store or your checkout; the one who closes the order and collects payment is always you, and intusell isn't a party to payment.
• It is not a managed "notify me when back in stock" list. It's not a system that announces a sold-out size to everyone at once; it's an assistant that makes an opportunistic re-offer to suitable WhatsApp customers being followed up.

In short: not a bot that spits out an answer with a single question, but a sales assistant that represents your boutique to the extent you train it — but never invents what it doesn't know.

Frequently asked questions

How long does Instagram boutique AI training take?

A working setup is completed within half a day: uploading product/size/shipping information to the knowledge base, setting the persona and tone, a few response rules, connecting WhatsApp and Instagram DM. The real refinement builds up in the first weeks as you approve and correct responses in the label review queue. Training is not one-off but continuous.

Is technical knowledge required for AI training?

No. All training is done from the panel; no code, API key, or developer is needed. You upload your product list (Excel/CSV), your size chart, and frequently asked questions to the knowledge base, choose the persona, and write the response rules in plain language. You connect Instagram with one-click OAuth and WhatsApp via QR from your phone.

Does the AI give wrong or made-up size/stock/price information?

No. The AI only uses the information you upload to the knowledge base; it does not invent a size, stock status, or price it is unsure of. When it does not know, it says so honestly and hands the chat to the team. This is the behavior that earns the most trust in boutique sales.

Does it automatically answer the "price?" questions in Instagram comments too?

WhatsApp and Instagram DM automation are approved and usable right away. Instagram comment automation ("price?", "is it in stock?" comments under a post) is rolled out gradually depending on Meta approval. You can automate DMs first and open comments when approval arrives; the two channel gates are independent. For details, see the Instagram and WhatsApp sales automation article.

Can the AI answer size, stock, and shipping questions and move them to an order?

Yes. When you upload the size chart, shipping times, product details, and payment methods to the knowledge base, the AI answers these repeated questions automatically and goes for the close at the right moment. In a situation it is unsure of, it does not add commentary; it hands off to the team. For the full flow of selling from DMs, take a look at the selling from DMs automation article.

Is my data safe under KVKK?

Yes. intusell is end-to-end encrypted and KVKK/GDPR compliant. PII (personal data) masking and explicit consent are applied to uploaded voice recordings; customer data is kept isolated per tenant.

Next step

You've trained your assistant; next is using it in day-to-day operations. The next article in the series explains how to run the intusell you trained on a real workday: an incoming size question, the "price?" downpour, cart/DM recovery, shipment tracking, and handoff to the team. Continue straight from there: how an Instagram boutique uses intusell. For the details of channel setup, the Instagram and WhatsApp sales automation article is ready, and for the finer points of selling from DMs, the selling from DMs automation article.

This guide was written for Instagram boutiques, but the same 5-step logic works in every sector. You can also take a look at the main articles of neighboring sectors: AI training for e-commerce, AI training for dental clinics, AI training for travel agencies, and AI training for clinics. If you'd like to see the setup specific to your boutique live, let's open your panel together in a 20-minute session with Get a demo, or write to hello@intusell.com. You can find all of the Instagram boutique capabilities on the Instagram boutiques solution and solutions pages, the package and quota details on the pricing page, and for other guides, take a look at the all articles list.