基于SpringAI的电商客服智能体

16Miku2025-05-19 22:28

简介

参考了spring-ai-alibaba-examples仓库中的playground-flight-booking项目，制作了一个基于SpringAI的具备RAG检索增强生成，对话记忆，函数调用，MCP服务调用等功能的电商客服智能体。

后端代码链接：https://github.com/16Miku/spring-ai-alibaba-agent

前端代码链接：https://github.com/16Miku/chat2-spring-ai

函数调用功能演示

成功读取到服务条款文档.

询问AI客服有哪些功能.

调用"创建新用户"功能

创建一个名为Miku的用户

日志信息中可以看到成功执行sql语句.

数据库中也可以看到新用户Miku的信息.

调用"更新用户信息"功能

将id为21的用户的密码修改为123456

日志信息中可以看到成功执行sql语句.

数据库中也可以看到id为21的用户的密码已修改为123456.

调用"根据用户ID获取用户信息"功能

查询id为1的用户的信息

日志信息中可以看到成功执行sql语句.

AI查询的数据符合数据库记录.???

调用"根据商品ID获取商品信息"功能

日志信息中可以看到成功执行sql语句.

AI查询的数据符合数据库记录.???

调用"创建订单"功能

为id为1的用户下单两件id为1的商品

log 复制代码 
            : Tool 'createOrder' called with request: createOrderRequest[userId=1, items=[OrderItem(id=null, orderId=null, itemId=1, itemName=null, itemPrice=null, quantity=2, createTime=null, updateTime=null)]]
2025-05-12T20:50:41.198+08:00  INFO 20112 --- [spring-ai-alibaba-agent] [tor-http-nio-12] c.m.s.service.impl.UserServiceImpl       : Fetching user with ID: 1
Creating a new SqlSession
Registering transaction synchronization for SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4509b93e]
JDBC Connection [HikariProxyConnection@50792400 wrapping com.mysql.cj.jdbc.ConnectionImpl@6d9bc4b3] will be managed by Spring
==>  Preparing: SELECT id, username, password, email, create_time, update_time FROM users WHERE id = ?
==> Parameters: 1(Long)
<==    Columns: id, username, password, email, create_time, update_time
<==        Row: 1, updatedTestUser, newSecurePassword456, [email protected], 2025-05-11 16:23:24, 2025-05-12 14:43:05
<==      Total: 1
Releasing transactional SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4509b93e]
2025-05-12T20:50:41.202+08:00  INFO 20112 --- [spring-ai-alibaba-agent] [tor-http-nio-12] c.m.s.service.impl.ItemServiceImpl       : Fetching item with ID: 1
Fetched SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4509b93e] from current transaction
==>  Preparing: SELECT id, name, price, stock, create_time, update_time FROM items WHERE id = ?
==> Parameters: 1(Long)
<==    Columns: id, name, price, stock, create_time, update_time
<==        Row: 1, 笔记本电脑, 5999.00, 97, 2025-05-11 16:23:24, 2025-05-12 14:43:12
<==      Total: 1
Releasing transactional SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4509b93e]
Fetched SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4509b93e] from current transaction
==>  Preparing: INSERT INTO orders (user_id, total_amount, status) VALUES (?, ?, ?)
==> Parameters: 1(Long), 11998.00(BigDecimal), PENDING(String)
<==    Updates: 1
Releasing transactional SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4509b93e]
2025-05-12T20:50:41.211+08:00  INFO 20112 --- [spring-ai-alibaba-agent] [tor-http-nio-12] c.m.s.service.impl.OrderServiceImpl      : Order created successfully with ID: 5
Fetched SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4509b93e] from current transaction
==>  Preparing: INSERT INTO order_items (order_id, item_id, item_name, item_price, quantity) VALUES (?, ?, ?, ?, ?)
==> Parameters: 5(Long), 1(Long), 笔记本电脑(String), 5999.00(BigDecimal), 2(Integer)
<==    Updates: 1
Releasing transactional SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4509b93e]
2025-05-12T20:50:41.225+08:00  INFO 20112 --- [spring-ai-alibaba-agent] [tor-http-nio-12] c.m.s.service.impl.OrderServiceImpl      : Order items inserted successfully for order ID: 5
2025-05-12T20:50:41.225+08:00  INFO 20112 --- [spring-ai-alibaba-agent] [tor-http-nio-12] c.m.s.service.impl.ItemServiceImpl       : Decreasing stock for item ID 1 by 2
Fetched SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4509b93e] from current transaction
==>  Preparing: UPDATE items SET stock = stock - ? WHERE id = ? AND stock >= ?
==> Parameters: 2(Integer), 1(Long), 2(Integer)
<==    Updates: 1
Releasing transactional SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4509b93e]
2025-05-12T20:50:41.229+08:00  INFO 20112 --- [spring-ai-alibaba-agent] [tor-http-nio-12] c.m.s.service.impl.ItemServiceImpl       : Successfully decreased stock for item ID 1
2025-05-12T20:50:41.230+08:00  INFO 20112 --- [spring-ai-alibaba-agent] [tor-http-nio-12] c.m.s.service.impl.OrderServiceImpl      : Stock decreased successfully for items in order ID: 5
Transaction synchronization committing SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4509b93e]
Transaction synchronization deregistering SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4509b93e]
Transaction synchronization closing SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4509b93e]
2025-05-12T20:50:41.233+08:00  INFO 20112 --- [spring-ai-alibaba-agent] [tor-http-nio-12] c.m.s.service.CustomTools                : Tool 'createOrder' response: createOrderResponse[order=Order(id=5, userId=1, totalAmount=11998.00, status=PENDING, createTime=null, updateTime=null, orderItems=[OrderItem(id=null, orderId=5, itemId=1, itemName=笔记本电脑, itemPrice=5999.00, quantity=2, createTime=null, updateTime=null)])]

日志信息中可以看到成功执行sql语句.

数据库中也可以看到新订单的信息.

调用"根据订单ID获取订单详情"功能

查询id为5的订单的信息

log 复制代码 
Tool 'getOrderById' called with request: getOrderByIdRequest[id=5]
2025-05-12T21:01:13.744+08:00  INFO 20112 --- [spring-ai-alibaba-agent] [tor-http-nio-14] c.m.s.service.impl.OrderServiceImpl      : Fetching order with ID: 5
Creating a new SqlSession
SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4a1efd6c] was not registered for synchronization because synchronization is not active
JDBC Connection [HikariProxyConnection@1544787482 wrapping com.mysql.cj.jdbc.ConnectionImpl@1b0cc4ad] will not be managed by Spring
==>  Preparing: SELECT id, user_id, total_amount, status, create_time, update_time FROM orders WHERE id = ?
==> Parameters: 5(Long)
<==    Columns: id, user_id, total_amount, status, create_time, update_time
<==        Row: 5, 1, 11998.00, PENDING, 2025-05-12 20:50:41, 2025-05-12 20:50:41
<==      Total: 1
Closing non transactional SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4a1efd6c]
Creating a new SqlSession
SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@70baae07] was not registered for synchronization because synchronization is not active
JDBC Connection [HikariProxyConnection@819545420 wrapping com.mysql.cj.jdbc.ConnectionImpl@1b0cc4ad] will not be managed by Spring
==>  Preparing: SELECT id, order_id, item_id, item_name, item_price, quantity, create_time, update_time FROM order_items WHERE order_id = ?
==> Parameters: 5(Long)
<==    Columns: id, order_id, item_id, item_name, item_price, quantity, create_time, update_time
<==        Row: 6, 5, 1, 笔记本电脑, 5999.00, 2, 2025-05-12 20:50:41, 2025-05-12 20:50:41
<==      Total: 1
Closing non transactional SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@70baae07]
2025-05-12T21:01:13.761+08:00  INFO 20112 --- [spring-ai-alibaba-agent] [tor-http-nio-14] c.m.s.service.CustomTools                : Tool 'getOrderById' response: getOrderByIdResponse[order=Order(id=5, userId=1, totalAmount=11998.00, status=PENDING, createTime=Mon May 12 20:50:41 CST 2025, updateTime=Mon May 12 20:50:41 CST 2025, orderItems=[OrderItem(id=6, orderId=5, itemId=1, itemName=笔记本电脑, itemPrice=5999.00, quantity=2, createTime=Mon May 12 20:50:41 CST 2025, updateTime=Mon May 12 20:50:41 CST 2025)])]

日志信息中可以看到成功执行sql语句.

调用"更新订单状态"功能

将id为5的订单改为已支付

日志信息中可以看到成功执行sql语句.

数据库中也可以看到订单状态由PENDING变为了PAID的信息.

调用"创建支付记录"功能

为id为5的订单创建支付记录

log 复制代码 
  : Tool 'createPayment' called with request: createPaymentRequest[orderId=5, amount=11998.0, paymentMethod=信用卡]
2025-05-12T21:12:26.736+08:00  INFO 20112 --- [spring-ai-alibaba-agent] [tor-http-nio-16] c.m.s.service.impl.OrderServiceImpl      : Fetching order with ID: 5
Creating a new SqlSession
Registering transaction synchronization for SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4e692385]
JDBC Connection [HikariProxyConnection@697353375 wrapping com.mysql.cj.jdbc.ConnectionImpl@1b0cc4ad] will be managed by Spring
==>  Preparing: SELECT id, user_id, total_amount, status, create_time, update_time FROM orders WHERE id = ?
==> Parameters: 5(Long)
<==    Columns: id, user_id, total_amount, status, create_time, update_time
<==        Row: 5, 1, 11998.00, PAID, 2025-05-12 20:50:41, 2025-05-12 21:06:35
<==      Total: 1
Releasing transactional SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4e692385]
Fetched SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4e692385] from current transaction
==>  Preparing: SELECT id, order_id, item_id, item_name, item_price, quantity, create_time, update_time FROM order_items WHERE order_id = ?
==> Parameters: 5(Long)
<==    Columns: id, order_id, item_id, item_name, item_price, quantity, create_time, update_time
<==        Row: 6, 5, 1, 笔记本电脑, 5999.00, 2, 2025-05-12 20:50:41, 2025-05-12 20:50:41
<==      Total: 1
Releasing transactional SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4e692385]
Fetched SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4e692385] from current transaction
==>  Preparing: SELECT id, order_id, amount, status, payment_method, transaction_id, create_time, update_time FROM payments WHERE order_id = ?
==> Parameters: 5(Long)
<==      Total: 0
Releasing transactional SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4e692385]
Fetched SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4e692385] from current transaction
==>  Preparing: INSERT INTO payments (order_id, amount, status, payment_method, transaction_id) VALUES (?, ?, ?, ?, ?)
==> Parameters: 5(Long), 11998.0(BigDecimal), PENDING(String), 信用卡(String), null
<==    Updates: 1
Releasing transactional SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4e692385]
2025-05-12T21:12:26.760+08:00  INFO 20112 --- [spring-ai-alibaba-agent] [tor-http-nio-16] c.m.s.service.impl.PayServiceImpl        : Payment record created successfully for order ID 5. Payment ID: 4
Transaction synchronization committing SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4e692385]
Transaction synchronization deregistering SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4e692385]
Transaction synchronization closing SqlSession [org.apache.ibatis.session.defaults.DefaultSqlSession@4e692385]
2025-05-12T21:12:26.763+08:00  INFO 20112 --- [spring-ai-alibaba-agent] [tor-http-nio-16] c.m.s.service.CustomTools                : Tool 'createPayment' response: createPaymentResponse[payment=Payment(id=4, orderId=5, amount=11998.0, status=PENDING, paymentMethod=信用卡, transactionId=null, createTime=null, updateTime=null)]

日志信息中可以看到成功执行sql语句.

数据库中也可以看到新支付单的信息.

调用"查询支付记录"功能

查询id为2的订单的支付单信息

日志信息中可以看到成功执行sql语句.

调用"更新支付记录"功能

将id为4的支付单状态改为已支付

日志信息中可以看到成功执行sql语句.

MySQL-MCP功能演示

使用的mcp-mysql-server链接：https://github.com/f4ww4z/mcp-mysql-server

connect_db

list_tables

query

execute

describe_table

总结

本文详细展示了如何基于 Spring AI 框架，结合 RAG、对话记忆、函数调用以及 MCP 服务调用等技术，构建一个功能丰富的电商客服智能体。

通过 Spring AI 提供的强大能力，我们能够：

  1. 利用 RAG 增强智能体的知识库，使其能够回答基于特定文档（如服务条款）的复杂问题。
  2. 通过 对话记忆 保持多轮对话的连贯性，提供更自然的交互体验。
  3. 实现 函数调用，让智能体能够执行实际的业务操作，例如创建用户、更新订单状态、查询商品信息等，将 AI 能力与后端服务紧密集成。
  4. 探索 MCP 服务调用，展示了智能体与数据库等外部系统的直接交互能力。

这个项目不仅验证了 Spring AI 在构建复杂智能体应用方面的潜力，也为电商领域提供了一个智能化的客服解决方案原型，能够有效提升用户体验和运营效率。未来的工作可以进一步优化模型的理解和决策能力，扩展更多业务功能，并提升系统的稳定性和可扩展性。

相关推荐
pjx9872 小时前
服务间的“握手”：OpenFeign声明式调用与客户端负载均衡
java·运维·spring·负载均衡
在未来等你3 小时前
互联网大厂Java求职面试：AI与大模型应用集成及云原生挑战
java·微服务·ai·kubernetes·大模型·embedding·spring ai
okok__TXF4 小时前
SpringBoot3+AI
java·人工智能·spring
找不到、了7 小时前
Spring-Beans的生命周期的介绍
java·开发语言·spring
caihuayuan48 小时前
React Native 0.68 安装react-native-picker报错：找不到compile
java·大数据·sql·spring·课程设计
不再幻想，脚踏实地9 小时前
Spring AOP从0到1
java·后端·spring
会敲键盘的猕猴桃很大胆9 小时前
Day11-苍穹外卖(数据统计篇)
java·spring boot·后端·spring·信息可视化
极客智谷9 小时前
Spring Cloud动态配置刷新：@RefreshScope与@Component的协同机制解析
后端·spring·spring cloud
Lizhihao＿10 小时前
Spring MVC 接口的访问方法如何设置
java·后端·spring·mvc
热门推荐
01从零安装 LLaMA-Factory 微调 Qwen 大模型成功及所有的坑02KGG转MP3工具|非KGM文件|解密音频03YOLOv8入门 | 重要性能衡量指标、训练结果评价及分析及影响mAP的因素【发论文关注的指标】04【SpeedAI科研小助手】2分钟极速解决知网维普重复率、AIGC率过高，一键全文降！文件格式不变，公式都保留的！05组基轨迹建模 GBTM的介绍与实现（Stata 或 R）06Coze扣子平台完整体验和实践（附国内和国际版对比）07Ubuntu24.04安装中文输入法08YOLOv5改进 | 添加CA注意力机制 + 增加预测层 + 更换损失函数之GIoU09DeepSeek各版本说明与优缺点分析10yolov8实战第五天——yolov8+ffmpg实时视频流检测并进行实时推流——（推流，保姆教学）