readme.md

Power Chatbot Code Component - powered by Open AI

The Power Chatbot component allows you to provide a truly conversational user experience/interface inside canvas apps and custom pages.

🤖 Glimpse of PowerChatbot

Glimpse of PowerChatbot configured inside Model driven app using custom page.

📘 How it works

Power Chatbot component solely rely on Custom connector, as it does most of heavy lifting, i.e. to interact with OpenAI API and bring back results.

👨🏻‍💻 Setup & 🪛 Configuration

Setup and configuration requires below mentioned steps.

Step 1 - Setup custom connector

Setup custom connector

Step 2 - Import code component

1. Enable Code components for canvas apps in your environment. See Code components for canvas apps - Power Apps | Microsoft Learn
2. Import the PowerChatbot solution into your environment.
3. Inside a solution, create a new custom page or canvas app.
4. From the side panel Insert menu, select Get more components from the footer section.
5. From the Code tab that appears inside the Import Components panel that appears, select the Power Chatbot component, and select Import.
6. From the side bar Insert menu, expand Code components, and select the Power Chatbot to place it on your canvas or custom page.
7. From the side panel Data menu, select + Add data and choose ChatGPTConnector custom connector created in previous step - Step 1.

The Power Chatbot renders the Item collection provided via Items property and expects it to be in a specific defined format.

Step 3 - Configure control

Power Chatbot triggers an OnSubmit event each time a text is submited either via click or Send Button.

1. Inside OnSubmit add the following

Set(
    varResponses,
    ChatGPTConnector.ChatAPI(
        "gpt-3.5-turbo",
        Collect(
            colChatMessages,
            {
                role: "user",
                content: Self.SubmittedText
            }
        )
    )
);
Collect(
    colChatMessages,
    First(varResponses.choices).message
);

2. Add colChatMessages to Items property.

Below is an example of a resulting format.

 Collect(
  colChatMessages,
  Table(
      {
          role: "user",
          content: "Hi"
      },
      {
          role: "assistant",
          content: "Hello! How can I assist you today?"
      },
      {
          role: "user",
          content: "I need help with Powerapps container component"
      },
      {
          role: "assistant",
          content: "Of course! I'd be happy to help you with PowerApps container component. What specifically do you need assistance with?"
      }
  )
);

🪂 Theming

Power Chatbot by default comes with a default theming. However if Use platform theme is turned on, it will inherit and rely on Modern controls and theme.

To enable this feature, see Enable modern controls and themes for your app

Important : Modern controls and themes are preview features and not meant for production use.

🪩 Seamless Experience

Power Chatbot uses Fluent UI React v9 framework along with Team's React Chat Package to bring the chat experience close to Microsoft Teams.

🔩 Configurability for better control

Power Chatbot comes with different configurable options to choose from inorder to change the user experience.

🛣️ Future roadmap

• UI improvements close to perfection. Especially from accessibility stand point.
• Explore the possibilities with Azure Open AI API and guidance.