Skip to main content

Get B2B Inquiries Directly with AI: Laifaxin Usage Guide

Laifaxin AI Path · Step 2

This guide shows you how to give the Laifaxin B2B Inquiry-Generation System to ZCode. It does not merely deliver a buyer list. It helps you find qualified prospects, build outreach and follow-up, generate inquiries, and reuse high-quality inquiry websites to find more similar prospects.

Before You Start

Complete all five requirements before running the inquiry workflow:

RequirementWhat you need to completeGuide
ZCodeInstall it, select “Connect BigModel to continue,” and sign inInstall and Sign In to ZCode
ZCode settingsEnable Full Access, Computer Control, and Browser Control; GLM-5.3-Flash is enough for most daily tasksRecommended ZCode Settings
Laifaxin SVIPPrepare an active SVIP account; the complete workflow in this guide requires SVIPRegister and Claim New-User Benefits · SVIP Benefits and Upgrade Options
Account connectionSign in to Laifaxin in your browser, retrieve the accesstoken, and let the AI test the connectionGet the accesstoken and Connect AI
Product informationPrepare the product name, target market, and main selling points; AI can help complete missing detailsUse the template below

If you already have an inquiry-customer website, an existing-customer website, or an ideal-customer website, prepare that as well. A real company website is the fastest seed for finding similar prospects. You can still begin with only a product.

The Complete Workflow in Two Diagrams

Choose the right starting point first, then enter the inquiry loop. If you already have a real customer website, do not restart by guessing audiences from the product—use that website to find similar prospects.

Diagram 1: Choose Your Starting Point

① Prerequisites
ZCode · Full Access · computer and browser controls · SVIP · accesstoken

② What do you have now?

⭐ Highest priority
An inquiry-customer website
Use the real inquiry customer's website as the seed

You have a proven example
A customer / ideal-customer website
Do not infer an audience again; use it directly as the seed

No customer website
Product information only
Analyze product → approve audience → find the first qualified prospect

↓ All three routes converge here
✅ A qualified customer-website seed

Diagram 2: Turn a Qualified Seed into Inquiries

① Qualified website seed
Start with a real customer
② Find similar + audit
Save only after the sample is accurate
③ Outreach + follow-up
Create emails and smart follow-up
④ Confirm activation
Revise first if not ready; sending starts only after confirmation
↓ After activation, follow the matching result

No inquiry yet
Revise the seed, audience, or emails, then return to Step ① and test again

✅ Qualified inquiry received
Apply the Inquiry tag → follow up one to one → use the strong inquiry website as a new seed

↺ Return the new seed to Step ① and find the next group of similar prospects

💡 The full loop: inquiry website / customer website / product-led audience inference → qualified website seed → similar prospects → outreach and follow-up → inquiry → a strong inquiry website becomes the next seed. If no inquiry arrives yet, review the seed, audience, and emails, then continue testing instead of treating one send as the end.

Core Result: Generate Inquiries Directly

Based on the team's current internal testing, among tasks where the product and audience fit, the list passes an audit, and sending is formally activated, about 80% receive at least one qualified inquiry on the activation day. A “qualified inquiry” means a buyer explicitly asks about pricing, samples, minimum order quantity, or purchasing cooperation. This is a current internal observation, not a guaranteed result for every task.

Step 1: Give It to ZCode

Create a new task in ZCode and paste the full prompt below. Replace XXX with your information. Add customer websites if you have them; otherwise leave those fields blank.

Please install and use this Laifaxin B2B inquiry-generation system:
https://github.com/tony-apan/laifaxin-b2b-sales

Follow these steps:
1. Download the repository.
2. Run python3 tools/onboard_check.py first.
3. Read README.md, SKILL.md, RULES.md, and output-templates/README.md in that order.
4. First check whether ZCode, Laifaxin SVIP, and the account connection are ready.
5. Choose the correct starting point from my information:
- If I have an inquiry-customer website, use it as the seed and find similar prospects.
- If I have an existing-customer or ideal-customer website, use it as the seed and find similar prospects.
- If I only have a product, analyze it, infer audiences, find the first qualified prospect, and then find similar prospects.
6. Guide me in plain language through prospect auditing, saving, outreach emails, and smart follow-up.
7. Do not activate sending until I explicitly say “Confirm activation.”
8. After an inquiry arrives, remind me to apply the “Inquiry” tag and reuse strong inquiry websites in the next similarity search.

My product: XXX
Target market: XXX
Main selling points: XXX
Existing inquiry-customer websites (optional): XXX
Existing customer or ideal-customer websites (optional): XXX

If ZCode cannot download the repository directly, open the GitHub toolkit, select Code → Download ZIP, extract it, open the folder in ZCode, and send the prompt above.

Step 2: Choose a Starting Point

Not everyone should begin with product analysis. First check whether you already have a real customer website:

What you already haveWhere to startWhat happens next
An inquiry-customer websiteHighest priority—use it directly as the seedFind similar prospects by website
An existing-customer or ideal-customer websiteUse it directly as the seedFind similar prospects by website
Only a productLet AI analyze the product and infer audiencesChoose one audience → find the first qualified prospect → find similar prospects

Start A: Existing Inquiry

Give the AI the website of a customer who recently sent a real inquiry. A company that has already inquired is often the strongest example of the customer type you want to reproduce.

This company has already sent us an inquiry: https://example.com
First explain why it fits our product, then use its website as a seed to find similar prospects.

Start B: Existing Customer Website

If you do not yet have an inquiry but already have past customers, priority accounts, or ideal customers, provide one to three websites.

These companies are ideal customers for us:
1. https://example-a.com
2. https://example-b.com

Compare their shared characteristics, select the strongest seed, and find similar prospects.

Start C: Product Only

If you do not have any customer website, continue in this order:

  1. analyze the product and market;
  2. infer four audiences that may buy it;
  3. recommend the best audience for the first test;
  4. ask you to approve that audience;
  5. find the first qualified prospect in that audience;
  6. use that prospect's website as the seed for finding similar prospects.

The key distinction: an audience is only a direction. The scalable workflow starts after the AI finds a qualified seed company and expands through similarity search.

Step 3: Find Similar Prospects

Every starting route eventually reaches the same step: use a qualified company website to find similar prospects at scale.

The AI finds candidates and audits a sample to determine whether they are likely to buy your product. Check whether:

  • the company's business is relevant;
  • it operates in the correct market;
  • it is a buyer, wholesaler, distributor, or target brand rather than an irrelevant company;
  • its website and contact information are usable;
  • its buying logic resembles that of the seed customer.

If the sample is inaccurate, do not rush to save it. Ask the AI to change the seed, adjust the audience, or refine the search direction, then audit again.

Step 4: Audit and Save

Only approve saving after the AI reports sample quality and expected volume:

StageWhat the AI doesWhat you do
Prospect auditSamples business relevance, market, and buying potentialDecide whether the direction is accurate
Volume approvalReports the estimated number of companies and contactsReply Confirm saving N
Save prospectsCreates tags and saves companies and contactsVerify the count and tags

💡 For the first run, test one audience. After the first replies arrive, decide whether to scale it or change direction.

Step 5: Build Outreach

After saving the prospects, the AI continues by:

  1. generating several outreach-email previews for the product and audience;
  2. producing the final emails after your edits;
  3. creating a multi-step smart follow-up plan;
  4. adding prospects to the correct plan;
  5. reporting the audience, tags, language, volume, and sending schedule.

You only need to check and approve:

  • whether the emails sound human;
  • whether the selling points are factually correct;
  • whether the audience, language, and market are correct;
  • whether the follow-up steps and intervals make sense.

Step 6: Confirm and Activate Sending

The AI should stop at the completion report by default and must not activate sending automatically. When everything is ready, reply:

Confirm activation

After activation, monitor the Laifaxin inbox and follow-up status. Replies or inquiries may arrive on the same day for well-matched products and audiences; products with longer buying cycles require more time and follow-up.

Step 7: Feed Inquiries Back

An inquiry is not the end of the workflow. It makes the next round more accurate:

  1. apply the Inquiry tag to stop later batch emails to that customer;
  2. handle needs, quotations, samples, and purchasing plans one to one;
  3. ask the AI which inquiry customers best match your target;
  4. extract the websites of those high-quality inquiry customers;
  5. use them as new seeds to find similar prospects again;
  6. repeat the audit, save, outreach, and follow-up stages.
The following customers have sent qualified inquiries:
1. https://inquiry-a.com
2. https://inquiry-b.com

Analyze their shared characteristics, choose the strongest seed to reproduce, and continue finding similar prospects.

This is the complete growth loop: real inquiries continuously improve your seeds, so future prospects increasingly resemble companies already willing to request information or pricing.

Product Input Example

We manufacture stainless steel insulated bottles for gift wholesalers, retail buyers, and private-label brands in the United States and Europe.
Our strengths are factory-direct pricing, custom logos, small trial orders, and fast sampling.
I do not have a customer website yet. First assess product fit, infer four audiences, and recommend the best one to test.

You can begin with incomplete information. The AI will organize a first product profile and ask you to confirm or complete it.

Update an Existing Installation

The repository changes over time. Do not create another empty installation. Send this prompt to ZCode:

Follow the “Update to the Latest Version” section in README.md and update my Laifaxin B2B inquiry-generation system.
Back up and preserve the local data in .local/ and runs/ first. After updating, run tools/onboard_check.py and report the health-check result.

See the latest release and release notes. When repository rules change, follow the current RULES.md and the relevant files under specs/.

Completion Checklist

  • ZCode is installed and signed in.
  • Full Access, Computer Control, and Browser Control are enabled.
  • An active Laifaxin SVIP account is ready.
  • Laifaxin is signed in and the accesstoken connection check passed.
  • You identified the correct route: inquiry website, customer website, or product only.
  • You obtained a qualified seed and found similar prospects.
  • You approved the prospect audit, save volume, outreach emails, and follow-up plan.
  • When ready to send, you explicitly replied “Confirm activation.”
  • After receiving an inquiry, you applied the tag and reused strong inquiry websites in the next similarity search.