The AgentivUse cases

Shopee buyer chats answered, even after hours

A Shopee chat agent that reads buyer messages through the Shopee Open Platform, answers product and order status questions from the shop's own listing and order data in Thai or English, sends refunds and complaints to staff on LINE, and only drafts replies until the owner trusts it

Function: Customer serviceEffort: About a weekPublished: 2026-10-23Market: Thailand

The problem

A Shopee shop in Thailand gets most buyer questions at night: does this come in size L, when will my order ship, is the two pack cheaper. The seller answers from a phone the next morning, by which time the buyer has bought elsewhere, and the ones who waited are already annoyed.

How the agent works

  1. It starts

    The shop authorises a partner app on the Shopee Open Platform (partner_id, partner_key, the shop authorisation code, then access and refresh tokens, the flow every community SDK on GitHub shows).

  2. The agent works

    On a short polling schedule the agent fetches new buyer messages per conversation (the Chat module's getMessage({ conversation_id }) in the shinryak wrapper), reads which item the buyer is asking about, and pulls the shop's own product and order data from the product and order endpoints (getItemList, getOrderDetail), then picks one of three actions. Product questions it answers from the listing, variation and stock data in Thai or English to match the buyer; order status questions it answers from the order record; refunds, complaints, bargaining and anything the data does not cover go to the staff LINE group with the conversation link and a suggested reply, and the agent tells the buyer a person will follow up. Replies go back through the IM send endpoint (/im/message/send, which per the MonveChen SDK accepts text, an image URL, an item_id or an order_id, so the agent can attach the product or order card). Every automated reply is logged with its conversation id, and for the first two weeks the agent only drafts: a staff member taps send in Seller Centre.

  3. The result

    The owner reviews the log weekly and adds the questions it got wrong to the shop's knowledge sheet in Google Sheets.

Worked example

The shinryak-shopee-api README on GitHub describes a wrapper for Shopee Open API v2 that supports the Shop API only (no Merchant API), with Product (incomplete), Shop, Order, Logistics, Chat, Public and Returns modules, and shows shop.Chat.getMessage({ conversation_id }) and shop.Order.getOrderDetail({ order_sn_list }); the repository is early stage with a single commit. The MonveChen/shopee-openapi-v2 README (TypeScript, Apache 2.0) says it covers only the IM part and documents /im/message/send with access_token, template_id and session_id plus optional txt, img_url, item_id and order_id, returning a message_id. The congminh1254/shopee-sdk README describes a TypeScript SDK with 29 managers such as Product, Order, Logistics and Payment that handles request signing and token refresh and needs Node.js 20 or later, but has no chat manager. None of the three reports a shop's results, so measure your own: the share of chats the agent answers alone versus hands to staff, and how many drafts staff edit during the trial.

Data protection. This use case touches personal data, so check the rules where you operate (for example GDPR, PDPA). Buyer messages and order records carry names, addresses, phone numbers and order history, personal data under Thailand's PDPA. Keep only the conversation id, the question type and the reply in the log, never copy addresses or phone numbers into the LINE group or the knowledge sheet, say in the shop's chat greeting that an automated assistant answers first, and check the Shopee Open Platform data rules and your AI vendor's processing terms before going live.

Join the Agent OS Club

8 weeks to the AI team a 50 person company would hire. Instant access to our agent blueprints. USD 15 a month, cancel any time.

Join the Agent OS Club

Building this for an organization? Learn more at theagentiv.com

Sources

Related use cases

First-line support answered around the clock

Intercom's Fin agent answers first-line support from your help centre, carries out routine actions such as refunds and account changes within your policies, and leaves the complex cases to your team

Customer serviceEffort: About a weekIntercom · Fin · Claude

A SaaS or online retail team in Singapore or Bangkok with two or three support staff gets the same questions after hours and across time zones: where is my order, how do I change my plan, can I get a refund. Each one waits for the morning shift, and the people who could fix the hard problems spend their day on the easy ones.

Every Shopee review answered, in your voice

An agent that reads new Shopee and Lazada reviews each morning, drafts a reply to each in the shop's tone and posts them after the owner approves

Customer serviceEffort: About a weekShopee · Lazada · ClaudeThailand

Reviews pile up across two seller centres and most go unanswered, while the one star review with a shipping complaint sits in public for days. Replying by hand means copying the same thank you note forty times and still missing the one that needed a real answer.

Hear the bad review before it goes public

An agent that asks each customer for a one tap rating on LINE after the job is done, thanks the happy ones and alerts the owner to the unhappy ones with a drafted reply

Customer serviceEffort: About a weekLINE · Google Sheets · ClaudeThailand

A clinic, salon, repair shop or delivery business only learns a customer was unhappy when the review appears on Google Maps or Facebook. Asking everyone by hand does not happen, and paper feedback forms end up in a drawer.