diff --git a/src/main/java/com/wecom/robot/config/RestTemplateConfig.java b/src/main/java/com/wecom/robot/config/RestTemplateConfig.java
new file mode 100644
index 0000000..a49fb63
--- /dev/null
+++ b/src/main/java/com/wecom/robot/config/RestTemplateConfig.java
@@ -0,0 +1,18 @@
+package com.wecom.robot.config;
+
+import org.springframework.context.annotation.Bean;
+import org.springframework.context.annotation.Configuration;
+import org.springframework.http.client.SimpleClientHttpRequestFactory;
+import org.springframework.web.client.RestTemplate;
+
+@Configuration
+public class RestTemplateConfig {
+
+    @Bean
+    public RestTemplate restTemplate() {
+        SimpleClientHttpRequestFactory factory = new SimpleClientHttpRequestFactory();
+        factory.setConnectTimeout(5000);
+        factory.setReadTimeout(5000);
+        return new RestTemplate(factory);
+    }
+}
diff --git a/src/main/java/com/wecom/robot/service/AiServiceClient.java b/src/main/java/com/wecom/robot/service/AiServiceClient.java
new file mode 100644
index 0000000..d1221f3
--- /dev/null
+++ b/src/main/java/com/wecom/robot/service/AiServiceClient.java
@@ -0,0 +1,13 @@
+package com.wecom.robot.service;
+
+import com.wecom.robot.dto.ai.ChatRequest;
+import com.wecom.robot.dto.ai.ChatResponse;
+
+import java.util.concurrent.CompletableFuture;
+
+public interface AiServiceClient {
+
+    CompletableFuture<ChatResponse> generateReply(ChatRequest request);
+
+    boolean healthCheck();
+}
diff --git a/src/main/java/com/wecom/robot/service/impl/AiServiceClientImpl.java b/src/main/java/com/wecom/robot/service/impl/AiServiceClientImpl.java
new file mode 100644
index 0000000..8193d82
--- /dev/null
+++ b/src/main/java/com/wecom/robot/service/impl/AiServiceClientImpl.java
@@ -0,0 +1,75 @@
+package com.wecom.robot.service.impl;
+
+import com.wecom.robot.config.AiServiceConfig;
+import com.wecom.robot.dto.ai.ChatRequest;
+import com.wecom.robot.dto.ai.ChatResponse;
+import com.wecom.robot.service.AiServiceClient;
+import io.github.resilience4j.circuitbreaker.annotation.CircuitBreaker;
+import io.github.resilience4j.timelimiter.annotation.TimeLimiter;
+import lombok.RequiredArgsConstructor;
+import lombok.extern.slf4j.Slf4j;
+import org.springframework.http.HttpEntity;
+import org.springframework.http.HttpHeaders;
+import org.springframework.http.MediaType;
+import org.springframework.http.ResponseEntity;
+import org.springframework.stereotype.Service;
+import org.springframework.web.client.RestTemplate;
+
+import java.util.concurrent.CompletableFuture;
+
+@Slf4j
+@Service
+@RequiredArgsConstructor
+public class AiServiceClientImpl implements AiServiceClient {
+
+    private static final String CHAT_ENDPOINT = "/ai/chat";
+    private static final String HEALTH_ENDPOINT = "/ai/health";
+
+    private final AiServiceConfig aiServiceConfig;
+    private final RestTemplate restTemplate;
+
+    @Override
+    @CircuitBreaker(name = "aiService", fallbackMethod = "generateReplyFallback")
+    @TimeLimiter(name = "aiService")
+    public CompletableFuture<ChatResponse> generateReply(ChatRequest request) {
+        log.info("[AC-MCA-04] 调用 AI 服务: sessionId={}", request.getSessionId());
+
+        String url = aiServiceConfig.getUrl() + CHAT_ENDPOINT;
+        HttpHeaders headers = new HttpHeaders();
+        headers.setContentType(MediaType.APPLICATION_JSON);
+        HttpEntity<ChatRequest> entity = new HttpEntity<>(request, headers);
+
+        ResponseEntity<ChatResponse> response = restTemplate.postForEntity(
+                url, entity, ChatResponse.class);
+
+        log.info("[AC-MCA-05] AI 服务响应: sessionId={}, shouldTransfer={}",
+                request.getSessionId(),
+                response.getBody() != null ? response.getBody().getShouldTransfer() : null);
+
+        return CompletableFuture.completedFuture(response.getBody());
+    }
+
+    public CompletableFuture<ChatResponse> generateReplyFallback(ChatRequest request, Throwable cause) {
+        log.warn("[AC-MCA-06][AC-MCA-07] AI 服务降级: sessionId={}, cause={}",
+                request.getSessionId(), cause.getMessage());
+
+        ChatResponse fallbackResponse = ChatResponse.fallbackWithTransfer(
+                "抱歉，我暂时无法回答您的问题，正在为您转接人工客服...",
+                cause.getMessage()
+        );
+
+        return CompletableFuture.completedFuture(fallbackResponse);
+    }
+
+    @Override
+    public boolean healthCheck() {
+        try {
+            String url = aiServiceConfig.getUrl() + HEALTH_ENDPOINT;
+            ResponseEntity<String> response = restTemplate.getForEntity(url, String.class);
+            return response.getStatusCode().is2xxSuccessful();
+        } catch (Exception e) {
+            log.error("[AC-MCA-04] AI 服务健康检查失败: {}", e.getMessage());
+            return false;
+        }
+    }
+}
diff --git a/src/main/resources/application.yml b/src/main/resources/application.yml
index 0eb8dbb..0d01b43 100644
--- a/src/main/resources/application.yml
+++ b/src/main/resources/application.yml
@@ -42,3 +42,17 @@ channel:
       enabled: false
     jd:
       enabled: false
+
+resilience4j:
+  circuitbreaker:
+    instances:
+      aiService:
+        failure-rate-threshold: 50
+        sliding-window-size: 10
+        sliding-window-type: COUNT_BASED
+        wait-duration-in-open-state: 30s
+        permitted-number-of-calls-in-half-open-state: 3
+  timelimiter:
+    instances:
+      aiService:
+        timeout-duration: 5s