How to Automate B2B Lead Collection and Follow-up Using AI
A practical system for capturing B2B leads from multiple channels, sorting them with AI, and running automated follow-up sequences. Designed for one person handling the volume of a small sales team.
AI Export Lab | August 3, 2026
Here is what a week looks like for a small exporter before automation:
A lead comes through the website contact form. Another through email. A third from a trade platform message. A fourth from a forwarded WhatsApp inquiry. Someone fills out a sample request. Someone sends a pricing question through LinkedIn. Each one lands in a different place. Each one requires a different response.
By Wednesday, you have eight unopened leads you meant to follow up on. By Friday, you cannot remember which ones you replied to and which ones you did not. Leads leak. Revenue leaks.
I fixed this with AI. The system is not complicated. It is three pieces working together: capture, sort, and follow-up. Here is the setup.
---
Step 1: Capture — All Leads in One Place
The first problem is that leads arrive scattered. Fix this by giving the AI a single inbox.
I route everything through a central collection point:
- Website contact forms → forward to a monitored email address
- Direct emails → CC the monitored address or forward automatically
- Trade platform messages → forward via platform settings
- WhatsApp, LinkedIn, etc. → manual forward takes 5 seconds, or use a forwarding integration
Now every lead sits in one inbox. AI can see all of it.
---
Step 2: Sort — AI Reads and Categories Every Lead
The AI reads each incoming message and writes three things into a tracking sheet:
- Lead type. "Quote request," "Sample request," "General inquiry," or "Spam." AI identifies this from the message content — product mentions, quantity requests, shipping terms.
- Priority score. High, medium, low. Rules are simple: contains an OE number = high. Requests pricing for a specific product = high. "Please send catalog" with no product mention = low.
- Suggested first action. "Send quote for AA-123 × 200 units." "Reply with product catalog PDF." "Ask for delivery address for sample."
The tracking sheet becomes your lead dashboard. You open it, see what needs action, and work from the top down. No more digging through five inboxes.
---
Step 3: Follow-up — Automated Sequences That Do Not Sound Automated
This is where most automation fails. A B2B buyer can spot a template email from the first sentence.
My approach: AI writes each follow-up fresh, using context from the original inquiry. It knows the product, the quantity, the timeline the buyer mentioned. It does not send "Dear valued customer, thank you for your inquiry." It sends "Hi, following up on your request for 200 units of AA-123 — our FOB Shanghai price is $12.50 per unit for that quantity. Production lead time is 15 days."
Follow-up cadence is automatic but human-reviewable:
- Day 1: Quote sent. AI flags it "awaiting reply."
- Day 3: If no reply, AI drafts a polite follow-up. I review and approve.
- Day 7: If still no reply, AI drafts a check-in with a value add — a spec sheet, a customer case, a shipping option the buyer might not have considered.
- Day 14: No more follow-up. Lead moves to "cold" list. AI sends a monthly digest to cold leads with new products or promotions.
Each follow-up is drafted by AI, reviewed by a human, and sent. The AI handles the remembering and the writing. I handle the final stamp of approval.
---
Materials on Demand: AI Generates Sales Documents
One thing that slowed me down was creating sales materials for each lead. A product spec sheet. A comparison chart against competitors. A shipping cost estimate table.
The AI now generates these from templates stored in a shared folder:
- Spec sheet. Enter the product code → AI pulls dimensions, materials, and compatibility data from the product database → generates a one-page PDF.
- Comparison table. Enter two product codes → AI creates a side-by-side spec comparison for the buyer.
- Quotation form. Enter product code and quantity → AI fills in pricing, MOQ, lead time, shipping terms into a pre-formatted template.
What used to take 20 minutes of copying and pasting between spreadsheets now takes a prompt: "Generate quote sheet for AA-123, 200 units, FOB Shanghai."
---
How to Build This Yourself
Start with step 1 only. Route all leads to one email address. That alone gives you visibility.
Add step 2 when you have more than five leads per day. Ask an AI: "Read these emails and sort them into a spreadsheet with lead type, priority, and recommended action."
Add step 3 when sorting is easy but following up still takes too much time. Ask the AI to draft follow-ups and schedule them in a calendar.
The whole system can run on the same API plan you use for everything else. The Volcengine Ark Agent Plan at ¥49/month handles lead sorting, follow-up drafting, and document generation with headroom to spare.
---
How are you handling lead follow-up? Curious what works for small export teams. Send me a message — I reply to all of these.