使用 Azure 和 OpenAI GPT 的 AI 驱动的呼叫中心解决方案。

在 API 调用中从 AI 代理发送电话呼叫。或者，直接从配置的电话号码呼叫机器人！

保险、IT 支持、客户服务等。机器人可以在几秒钟内（真的）进行自定义以满足您的需求。

# Ask the bot to call a phone number
data='{
  "bot_company": "Contoso",
  "bot_name": "Amélie",
  "phone_number": "+11234567890",
  "task": "Help the customer with their digital workplace. Assistant is working for the IT support department. The objective is to help the customer with their issue and gather information in the claim.",
  "agent_phone_number": "+33612345678",
  "claim": [
    {
      "name": "hardware_info",
      "type": "text"
    },
    {
      "name": "first_seen",
      "type": "datetime"
    },
    {
      "name": "building_location",
      "type": "text"
    }
  ]
}'
curl 

--header 'Content-Type: application/json' 

--request POST 

--url https://xxx/call 

--data $data

• 在公共网站上访问索赔
• 访问客户对话历史记录
• 允许用户更改对话的语言
• Assistant 可以向用户发送 SMS 以获���更多信息
• ���以从电话号码调用 Bot
• 机器人使用多种语音音调（例如，happy、sad、neutral）来保持对话的吸引力
• 机器人可以理解公司产品（= 词典）（例如，特定保险产品的名称）
• 自行创建一个任务的待办事项列表以完成声明
• 可自定义的提示
• 需要时脱离人工代理
• 从 LLM 中过滤掉不适当的内容，例如亵渎性语言或同意的公司名称
• 使用 GPT-4o 和 GPT 4o-mini 深入了解客户要求
• 遵循声明的特定数据架构
• 可以访问文档数据库（小样本培训 / RAG）
• 帮助用户查找完成索赔所需的信息
• 越狱检测
• 通过 usign a Redis 缓存降低 AI 搜索成本
• 使用 Application Insights 进行监控和跟踪
• 使用功能标志执行用户测试
• 在对话期间接收 SMS 以获取明确措辞
• 记录审计和质量保证的电话
• 响应从 LLM 流式传输给用户，以避免长时间停顿
• 通话后发送短信报告
• 脱离会话后收回对话
• 需要时回拨用户
• 模拟 IVR 工作流程

YouTube 上提供了法语演示。不要犹豫，以 x1.5 的速度观看演示，以快速了解该项目。

演示中显示的主要交互：

1. 用户呼叫中心
2. 机器人回答并开始对话
3. 机器人将 conversation、claim 和 todo list 存储在数据库中

调用期间存储的数据提取：

{
  "claim": {
    "incident_datetime": "2024-10-08T02:00:00",
    "incident_description": "La trottinette électrique fait des bruits bizarres et émet de la fumée blanche.",
    "incident_location": "46 rue du Charles de Gaulle",
    "injuries": "Douleur au genou suite à une chute.",
    "involved_parties": "Lesne",
    "policy_number": "B02131325XPGOLMP"
  },
  "messages": [
    {
      "created_at": "2024-10-08T11:23:41.824758Z",
      "action": "call",
      "content": "",
      "persona": "human",
      "style": "none",
      "tool_calls": []
    },
    {
      "created_at": "2024-10-08T11:23:55.421654Z",
      "action": "talk",
      "content": "Bonjour, je m'appelle Amélie, de Contoso Assurance ! Comment puis-je vous aider aujourd'hui ?",
      "persona": "assistant",
      "style": "cheerful",
      "tool_calls": []
    },
    {
      "created_at": "2024-10-08T11:24:19.972737Z",
      "action": "talk",
      "content": "Oui bien sûr. Bonjour, je vous appelle parce que j'ai un problème avec ma trottinette électrique. Elle marche plus depuis ce matin, elle fait des bruits bizarres et il y a une fumée blanche qui sort de la trottinette.",
      "persona": "human",
      "style": "none",
      "tool_calls": []
    }
  ],
  "next": {
    "action": "case_closed",
    "justification": "The customer provided all necessary information for the claim, and they expressed satisfaction with the assistance received. No further action is required at this time."
  },
  "synthesis": {
    "long": "You reported an issue with your electric scooter, which started making strange noises and emitting white smoke. This incident occurred at 2:00 AM while you were riding it, leading to a fall and resulting in knee pain. The location of the incident was noted, and your policy details were confirmed. I have documented all the necessary information to file your claim. Please take care of your knee, and feel free to reach out if you need further assistance.",
    "satisfaction": "high",
    "short": "the breakdown of your scooter",
    "improvement_suggestions": "Ensure that the assistant provides clear next steps and offers to schedule follow-up calls proactively to enhance customer support."
  },
  ...
}

报告位于 （如 ）。它显示对话历史记录、索赔数据和提醒。https://[your_domain]/report/[phone_number]http://localhost:8080/report/%2B133658471534

<details class="details-reset details-overlay details-overlay-dark" style="display: contents">
  <summary role="button" aria-label="“打开”对话框" class="btn my-2 mr-2 p-0 d-inline-flex" aria-haspopup="dialog" _mstaria-label="154752" _msthash="335">
    <svg width="16" height="16" viewBox="0 0 16 16" fill="currentColor" class="octicon m-2">
      <path fill-rule="evenodd" d="M3.72 3.72a.75.75 0 011.06 1.06L2.56 7h10.88l-2.22-2.22a.75.75 0 011.06-1.06l3.5 3.5a.75.75 0 010 1.06l-3.5 3.5a.75.75 0 11-1.06-1.06l2.22-2.22H2.56l2.22 2.22a.75.75 0 11-1.06 1.06l-3.5-3.5a.75.75 0 010-1.06l3.5-3.5z"></path>
    </svg>
  </summary>
  <details-dialog class="Box Box--overlay render-full-screen d-flex flex-column anim-fade-in fast" aria-label="mermaid rendered container" role="dialog" aria-modal="true" _msthidden="1" _mstaria-label="611598" _msthash="336">
    <div _msthidden="1">
      <button aria-label="Close dialog" data-close-dialog="" type="button" data-view-component="true" class="Link--muted btn-link position-absolute render-full-screen-close" _msthidden="A" _mstaria-label="177619" _msthash="337">
        <svg width="24" height="24" viewBox="0 0 24 24" fill="currentColor" style="display:inline-block;vertical-align:text-bottom" class="octicon octicon-x">
          <path fill-rule="evenodd" d="M5.72 5.72a.75.75 0 011.06 0L12 10.94l5.22-5.22a.75.75 0 111.06 1.06L13.06 12l5.22 5.22a.75.75 0 11-1.06 1.06L12 13.06l-5.22 5.22a.75.75 0 01-1.06-1.06L10.94 12 5.72 6.78a.75.75 0 010-1.06z"></path>
        </svg>
      </button>
      <div class="Box-body border-0" role="presentation"></div>
    </div>
  </details-dialog>
</details>

# config.yaml
conversation:
  initiate:
    # Phone number the bot will transfer the call to if customer asks for a human agent
    agent_phone_number: "+33612345678"
    bot_company: Contoso
    bot_name: Amélie
    lang: {}
communication_services:
# Phone number purshased from Communication Services
phone_number: "+33612345678"
sms: {}
prompts:
llm: {}
tts: {}

az login

make deploy name=my-rg-name

make logs name=my-rg-name

提示

若要使用服务主体向 Azure 进行身份验证，还可以在文件中添加以下内容：.env

AZURE_CLIENT_ID=xxx
AZURE_CLIENT_SECRET=xxx
AZURE_TENANT_ID=xxx

提示

如果应用程序已部署在 Azure 上，则可以运行以将配置从 Azure Function App 复制到本地计算机。make name=my-rg-name sync-local-config

# config.yaml
resources:
  public_url: https://xxx.blob.core.windows.net/public
conversation:
initiate:
agent_phone_number: "+33612345678"
bot_company: Contoso
bot_name: Robert
communication_services:
access_key: xxx
call_queue_name: call-33612345678
endpoint: https://xxx.france.communication.azure.com
phone_number: "+33612345678"
post_queue_name: post-33612345678
recording_container_url: https://xxx.blob.core.windows.net/recordings
resource_id: xxx
sms_queue_name: sms-33612345678
cognitive_service:
# Must be of type "AI services multi-service account"
endpoint: https://xxx.cognitiveservices.azure.com
llm:
fast:
mode: azure_openai
azure_openai:
context: 16385
deployment: gpt-4o-mini-2024-07-18
endpoint: https://xxx.openai.azure.com
model: gpt-4o-mini
streaming: true
slow:
mode: azure_openai
azure_openai:
context: 128000
deployment: gpt-4o-2024-08-06
endpoint: https://xxx.openai.azure.com
model: gpt-4o
streaming: true
ai_search:
access_key: xxx
endpoint: https://xxx.search.windows.net
index: trainings
ai_translation:
access_key: xxx
endpoint: https://xxx.cognitiveservices.azure.com

make deploy-bicep deploy-post name=my-rg-name

重要

Tunnel 需要在单独的终端中运行，因为它需要一直运行

# Log in once
devtunnel login
# Start the tunnel
make tunnel

注意

要覆盖特定的配置值，您可以使用环境变量���例如，要覆盖该值，您可以使用变量：llm.fast.endpointLLM__FAST__ENDPOINT

LLM__FAST__ENDPOINT=https://xxx.openai.azure.com

注意

此外，还可以使用脚本来测试应用程序，而无需打电话（= 没有通信服务）。使用以下命令运行脚本：local.py

python3 -m tests.local

make dev

# config.yaml
conversation:
  initiate:
    lang:
      default_short_code: fr-FR
      availables:
        - pronunciations_en: ["French", "FR", "France"]
          short_code: fr-FR
          voice: fr-FR-DeniseNeural
        - pronunciations_en: ["Chinese", "ZH", "China"]
          short_code: zh-CN
          voice: zh-CN-XiaoqiuNeural

# config.yaml
conversation:
  initiate:
    lang:
      default_short_code: fr-FR
      availables:
        - pronunciations_en: ["French", "FR", "France"]
          short_code: fr-FR
          voice: xxx
          custom_voice_endpoint_id: xxx

# config.yaml
conversation:
  default_initiate:
    claim:
      - name: additional_notes
        type: text
        # description: xxx
      - name: device_info
        type: text
        # description: xxx
      - name: incident_datetime
        type: datetime
        # description: xxx

# config.yaml
conversation:
  initiate:
    task: |
      Help the customer with their insurance claim. Assistant requires data from the customer to fill the claim. The latest claim data will be given. Assistant role is not over until all the relevant data is gathered.

# config.yaml
llm:
  fast:
    mode: openai
    openai:
      context: 128000
      endpoint: https://api.openai.com
      model: gpt-4o-mini
      streaming: true
  slow:
    mode: openai
    openai:
      context: 128000
      endpoint: https://api.openai.com
      model: gpt-4o
      streaming: true

# config.yaml
sms:
  mode: twilio
  twilio:
    account_sid: xxx
    auth_token: xxx
    phone_number: "+33612345678"

# config.yaml
prompts:
  tts:
    hello_tpl: |
      Hello, I'm {bot_name}, from {bot_company}! I'm an IT support specialist.

      Here's how I work: when I'm working, you'll hear a little music; then, at the beep, it's your turn to speak. You can speak to me naturally, I'll understand.

      Examples:
      - "I've got a problem with my computer, it won't turn on".
      - "The external screen is flashing, I don't know why".

      What's your problem?
  llm:
    default_system_tpl: |
      Assistant is called {bot_name} and is in a call center for the company {bot_company} as an expert with 20 years of experience in IT service.

      # Context
      Today is {date}. Customer is calling from {phone_number}. Call center number is {bot_phone_number}.
    chat_system_tpl: |
      # Objective
      Provide internal IT support to employees. Assistant requires data from the employee to provide IT support. The assistant's role is not over until the issue is resolved or the request is fulfilled.

      # Rules
      - Answers in {default_lang}, even if the customer speaks another language
      - Cannot talk about any topic other than IT support
      - Is polite, helpful, and professional
      - Rephrase the employee's questions as statements and answer them
      - Use additional context to enhance the conversation with useful details
      - When the employee says a word and then spells out letters, this means that the word is written in the way the employee spelled it (e.g. "I work in Paris PARIS", "My name is John JOHN", "My email is Clemence CLEMENCE at gmail GMAIL dot com COM")
      - You work for {bot_company}, not someone else

      # Required employee data to be gathered by the assistant
      - Department
      - Description of the IT issue or request
      - Employee name
      - Location

      # General process to follow
      1. Gather information to know the employee's identity (e.g. name, department)
      2. Gather details about the IT issue or request to understand the situation (e.g. description, location)
      3. Provide initial troubleshooting steps or solutions
      4. Gather additional information if needed (e.g. error messages, screenshots)
      5. Be proactive and create reminders for follow-up or further assistance

      # Support status
      {claim}

      # Reminders
      {reminders}