How to Set Up Intent-Conditional Lead Capture on Your Shopify Store

The Problem With One-Size-Fits-All Email Capture

Most Shopify chat widgets treat every email capture the same way: customer gives email, widget gives coupon code, done. That works for discount seekers and newsletter signups. But what happens when a wholesale buyer asks about bulk pricing and gets a 10% off code instead of a callback from your sales team?

You waste a discount code. The buyer gets a confusing experience. And the lead sits in your email list alongside newsletter signups with no context about what they actually wanted.

How Intent Classification Works

Milly Chat's AI classifies every email capture into one of four intents:

• Promotional — Customer asked about coupons, promo codes, sales, or wants marketing emails. The only intent eligible for coupon generation.
• Business — Wholesale, B2B, partnerships, custom quotes, collaboration. Never generates a coupon.
• Support — Help requests, complaints, or order issues that need a human follow-up. Never generates a coupon.
• General — Catch-all for captures that don't fit the other three. Never generates a coupon.

This classification happens automatically. The AI reads the conversation context and decides. You don't need to write any rules for the default behavior—promotional captures get coupons, business and support inquiries don't.

Teaching the AI When to Collect More

The default flow captures email and name. But for B2B or collaboration inquiries, you probably want more context: what they want to use your product for, the scale of the project, their timeline.

This is where AI Instructions come in. Add a short rule describing the intent you care about, and the AI will route matching customers through a richer capture flow:

If the user is asking about wholesale or B2B pricing,
ask for their email, company name, and what they're
looking for. Pass the details to the body field of
capture_email so we can prioritize the follow-up.

The AI writes a 1-2 sentence summary of the customer's actual request into the lead's body field. When you open the lead in the dashboard, you see the summary alongside the email—so you can triage whether to follow up without reading the entire conversation.

What You See in the Leads Tab

Every captured lead shows up in the Leads page with search, date range filtering, and pagination. Click any row to open the slideout drawer:

• Page URL where the conversation happened (clickable, opens in a new tab)
• AI Summary of what the customer was asking about
• Conversation Preview showing the last 20 messages inline
• Full Replay Link that opens the conversation replay in a new tab
• Coupon Code badge (only for the promotional intent)

Setting It Up

1. Go to Configure → Lead Capture and enable email capture
2. Choose your lead platform (Milly Chat database, Maestra, Klaviyo, HubSpot, or Zendesk — Zendesk only fires for support-intent leads, so pair it with a different default platform)
3. Optionally enable Coupon Reward for promotional signups — this uses the separate Shopify Discount adapter and runs alongside whichever lead platform you picked
4. Go to Configure → Response → AI Instructions and add intent-specific rules for your business (wholesale, partnerships, custom quotes, etc.)
5. Optionally set up a Custom Webhook to push every lead to your own CRM, Slack, or internal tools

That's it. The AI handles the intent classification, coupon gating, body summary, and capture flow automatically. You configure what intents matter to your business, and the system does the rest.

Why This Matters

The difference between a discount seeker and a wholesale buyer is the difference between a $15 order and a $15,000 order. If your chat widget treats both the same, you're leaving the $15,000 conversation on the table—or worse, opening it with a coupon code that makes you look like you don't understand the customer's intent.

Intent-conditional lead capture solves this by letting the AI do what it's good at: reading the conversation, classifying the intent, collecting the right information, and routing the lead to the right place—all without the customer noticing any of the machinery underneath.