After-Sales Execution Playbook
How to configure logistics, returns, refunds, compensation, disputes, human review, and evidence capture in Aijia Customer Service.
Goal
This playbook helps cross-border ecommerce, marketplace, and social commerce teams configure after-sales support execution.
The goal is not to let AI replace every support judgment. The goal is to let AI handle repetitive verification, drafting, organization, preparation, and evidence capture while humans own exceptions, risk, and customer relationships.
After-Sales Intent Types
Start with seven intent groups:
- logistics: in transit, delayed, delivered, lost, address exception;
- returns: return window, category restrictions, shipping responsibility, evidence requirements;
- refund status: refund path, platform processing time, abnormal state;
- replacement: wrong item, missing item, damage, size or color issue;
- disputes: marketplace disputes, chargebacks, negative reviews, public complaints;
- campaign after-sales: coupons, price difference, gifts, campaign promises;
- unknown or high-risk: uncertain identity, policy conflict, strong emotion.
Each intent should have knowledge, permissions, review rules, and evidence requirements.
Recommended Workflow
- Capture customer message, channel, language, order clues, and customer history.
- Identify after-sales intent and risk level.
- Retrieve product, logistics, refund, regional, platform, and brand policies.
- If context is needed, check order, logistics, or refund status within authorized boundaries.
- Generate a reply draft, action suggestion, and evidence summary.
- Let agents confirm low-risk cases; route high-risk cases to supervisor review.
- Save final reply, screenshots, approvals, outcome, and case summary.
Review Rules
The following cases should usually require review before action:
- refunds, compensation, credits, or coupon commitments;
- disputed orders and marketplace complaints;
- public negative replies;
- product quality, restricted categories, or compliance risk;
- uncertain identity, order, or evidence;
- conflict between platform policy and brand policy;
- exceptions outside the standard after-sales policy.
Review rules should be configurable by store, region, role, and channel.
Evidence Requirements
Important after-sales cases should preserve:
- original customer message;
- retrieved knowledge and policy version;
- order, logistics, refund, or platform screenshots;
- AI action plan and reply draft;
- reviewer, approval note, and time;
- final sent message;
- outcome and follow-up tags.
Evidence supports both risk control and operational review.
Rollout Advice
Start with read-only and reviewed workflows:
- connect one high-volume after-sales channel;
- organize after-sales policy and risk boundaries;
- enable draft replies;
- enable read-only order and logistics verification;
- turn refunds, compensation, and disputes into reviewed actions;
- review 2-4 weeks of data before enabling low-risk automation.
