Configure pre-sales AI support across channels, product knowledge, discounts, stock, intent levels, back-office verification, human follow-up, and conversion metrics.

The goal of pre-sales AI support is to detect buying opportunities inside comments, DMs, marketplace inquiries, and private-domain messages, then respond accurately and follow up consistently.

1. Define Pre-Sales Channels

List the channels that generate purchase questions:

short-video comments and DMs;
live comments and shopping-cart inquiries;
marketplace product pages, order pages, and support entries;
WhatsApp, LINE, WeChat, WeCom, and email;
communities, influencer campaigns, and ad landing pages.

Each channel should have its own tone, length, call to action, and risk boundary.

2. Prepare Product and Campaign Knowledge

Pre-sales knowledge should cover:

product specs, size, material, color, and target users;
use cases, pairing advice, and common misunderstandings;
stock, pre-order, purchase limits, restock, and alternatives;
coupons, gifts, bundles, live prices, and marketplace campaigns;
delivery time, shipping fees, tax, and serviceable regions;
restricted promises around performance, effect, compatibility, or price protection.

Maintain knowledge by channel, store, country, language, and campaign period so outdated campaign rules are not reused.

3. Configure Intent Levels

Group messages into four levels:

Level	Signal	Handling
Low intent	casual browsing, emoji, simple likes	short reply or follow prompt
Medium intent	specs, price, campaign, delivery	auto-reply or draft
High intent	repeated confirmation, checkout, comparison	mark for follow-up, use human when needed
Risk intent	special promises, sensitive claims, public complaints	human review or takeover

Intent levels help the team spend human attention where conversion or risk is highest.

4. Define Back-Office Verification Scenarios

These pre-sales questions often require verification:

whether campaign stock is still available;
whether a coupon is usable or already claimed;
whether an order was created or payment failed;
whether the customer has prior orders or after-sales history;
whether the delivery promise applies to the customer's region;
whether the marketplace dashboard shows a product exception.

Verification results should become response evidence and be retained when needed.

5. Configure Human Follow-Up Queue

Route these customers into a human queue:

high-value or bulk orders;
repeated questions without checkout;
explicit purchase intent blocked by one issue;
special discount or delivery request;
public negative comments that affect conversion;
highly uncertain product fit.

The queue should show original message, AI judgment, referenced knowledge, back-office evidence, and suggested next step.

6. Risk Rules

Pre-sales support should block:

promising stock before verification;
promising refunds, compensation, or price protection before policy confirmation;
exaggerated claims around performance, health, return, or compatibility;
handling privacy, order, or dispute details in public comments;
bypassing platform rules.

High-risk messages should default to review-before-send.

7. Acceptance Metrics

Track these during a pilot:

first response time;
comment-to-DM conversion rate;
high-intent detection accuracy;
product knowledge hit rate;
discount and stock verification time;
conversion after human follow-up;
lost-deal reason attribution;
risky reply interception rate.

8. Weekly Review

Every week, review:

missing knowledge behind frequent questions;
replies frequently edited by humans;
expired discount or stock rules;
channels with low conversion;
risky wording that needs new block rules;
high-intent customers not followed up in time.

Pre-sales AI support is not only faster replies. It turns every buying signal into a managed operating process.

Sales Conversion Support Playbook