reddit · SideProject · 09-19 11:50 60 分 ← 返回列表

Every WhatsApp order bot I've used is a phone tree in disguise — so I built one that just understands you

Every WhatsApp order bot I've used is a phone tree in disguise so I built one that just understands you You know the drill. Reply 1 for menu. Reply 2 for locations. Reply 9 to scream into the void. It's a phone tree wearing a chat bubble. I wanted to see if I could skip all of that no keywords, no numbered replies, just talk to it like you'd talk to the shop's actual staff. So for a juice delivery shop I'm working with, I built a WhatsApp assistant that: Understands plain language (

五维评分

付费意愿

需求真实度

个人可落地

竞争空间

新鲜度

AI 分析

痛点：现有 WhatsApp 点单机器人本质是伪装的电话按键树（回复1看菜单、回复2看位置），用户被迫用关键词和数字回复，交互僵硬、购物车/支付/地址状态极易丢失。

目标用户：新兴市场（如肯尼亚）使用 WhatsApp + M-Pesa 接单的小微商家（果汁店、餐饮外卖、本地零售），以及为其做自动化的开发者/代理商。

证据：Every WhatsApp order bot I've used is a phone tree in disguise... It's a phone tree wearing a chat bubble. / so for a juice delivery shop I'm working with, I built a WhatsApp assistant that... / Fires an M-Pesa STK push for checkout, or falls back to pay-on-delivery... / Posts the finished order into the shop's actual backend with an order number, no human re-typing anything on their end

变现：按商家订阅的 SaaS（月费 + 消息量阶梯），或按成交订单抽取佣金；面向代理商的 white-label 授权。

MVP：用 WhatsApp Business API + LLM 做意图解析，接商家自身网站/表格同步商品目录，维护跨会话持久化的购物车状态，集成 M-Pesa STK Push 及货到付款回退，订单直接写入商家后台；先服务 1-2 家本地商家验证闭环。

竞品：Wati、Interakt、Zoko、ManyChat、Twilio WhatsApp API、WhatsApp 官方 Business App

摘要：开发者给肯尼亚一家果汁配送店做了 WhatsApp 点单助手，用自然语言（含 niaje/sasa 等斯瓦希里语混用）替代按键树，自动读目录、算运费、建购物车，并用 M-Pesa STK Push 结账、订单直写商家后台。作者认为真正难点不是 NLP 而是跨会话的状态管理（购物车、支付重试、地址解析）。有真实商家在配合，但帖子未披露价格或收入数据。