top of page

Power Virtual Agents: Building Enhanced Chatbots with GPT

This post is intended to start a series of articles on Artificial Intelligence applied to the Dynamics 365 platform. We start with a bit of context and theory to understand what others will bring us!!

Power Virtual Agents is a chatbot platform that uses Natural Language Processing (NLP) technology to improve the ability of chatbots to understand and respond to customer requests. One of the PLN technologies used in Power Virtual Agents is the Generative Pre-trained Transformer (GPT) model.

What is GPT?

GPT is a machine learning model that relies on processing large amounts of data to understand natural language. This model

It is based on the Transformer, a neural network architecture that uses attention to process input. The GPT model was developed by OpenAI and is capable of generating high-quality and coherent text in different language processing tasks, such as machine translation, question answering, and text generation.

When used in Power Virtual Agents, GPT enables chatbots to better understand the meaning behind user requests and generate appropriate and consistent responses. Additionally, the GPT model adapts and improves over time as the chatbot is used, allowing for greater accuracy and quality of responses.

Other AI models

There are other AI models available in Azure OpenAI:

  1. GPT (Generative Pre-trained Transformer) is a natural language model that uses advanced deep learning techniques to generate coherent and natural text. This model has been trained with a large amount of text data and is capable of performing tasks such as text generation, machine translation, and answering questions.

  2. Da Vinci is an improved version of the GPT-3 model. It uses more advanced deep learning techniques to produce higher quality and more coherent texts than other natural language models. In addition, it is also capable of answering specific questions in a more precise and detailed way.

  3. Curie is another natural language model available in Azure OpenAI, which focuses on generating text for specific tasks. This model uses advanced deep learning techniques to generate text adapted to different writing styles and tones of voice. It is an ideal choice for writing emails, generating marketing copy, and writing product descriptions.

How to create a chatbot with Power Virtual Agents?

To create a chatbot with Power Virtual Agents, several steps must be followed. First, it is necessary to define the goals and topics of the chatbot to ensure that it is designed to meet the specific needs of the users. Next, the chatbot's conversation flow must be designed and responses configured using the visual editor or natural language.

The Power Virtual Agents visual editor allows you to design the conversation flow by dragging and dropping blocks of content and configuring the corresponding values. On the other hand, natural language allows you to configure responses using simple phrases in a natural language that the chatbot can understand.

It is important to test and debug the chatbot before publishing and distributing it through different channels. Power Virtual Agents offers testing and debugging tools to make sure the chatbot is working properly.

How to embed a chatbot in a website?

One of the advantages of Power Virtual Agents is that it allows chatbots to be embedded in websites and applications. To embed a chatbot on a website, the following steps must be followed:

  1. In the Power Virtual Agents portal, click the "Add to a website" button on the chatbot you want to embed.

  2. Select the "Website" channel and click "Next".

  3. Configure the appearance of the chatbot to match the design of the website.

  4. Select the type of authentication you want to use and click "Next".

  5. Copy the generated embed code and paste it into the website source code.

Power Virtual Agents offers the ability to customize the appearance of the chatbot to match the branding or design of the website or app. This means that chatbots can be an effective tool for interacting with customers and improving the user experience without disrupting the design or functionality of the website.

What are the preview conditions for Power Virtual Agents with GPT?

The preview version of Power Virtual Agents incorporating GPT is currently in a testing and experimentation phase. This preview version has certain conditions and limitations, such as:

  • The preview is available only in certain regions and selected languages.

  • The preview is only available to enterprise customers with a Dynamics 365 or Microsoft Power Platform license.

  • The preview version may have limitations in terms of scalability and customization.

  • The preview version may have bugs or performance issues.

It's important to be aware of these limitations when using the preview version of Power Virtual Agents with GPT, and be prepared to work with Microsoft to resolve any issues that may arise.

Differences between Bot Framework Composer and Power Virtual Agents web editor

Both the Bot Framework Composer and the Power Virtual Agents web editor are chatbot creation tools that allow users to create and design chatbots visually, without the need to write code.

Even though both tools are similar in some ways, there are significant differences between them, below are some of the key differences:

  • Platform: Bot Framework Composer is a desktop tool that runs on Windows, while Power Virtual Agents web editor runs in the cloud on the Microsoft Power Platform service.

  • Integration with third-party services: Bot Framework Composer integrates with a wide variety of third-party services, such as Azure Cognitive Services, Azure Business Logic, and RESTful Web Services, allowing users to build more sophisticated and custom bots. On the other hand, the Power Virtual Agents web editor focuses on integration with Microsoft services, such as Dynamics 365, Power Automate, and Power Apps.

  • Customization and control: The Bot Framework Composer gives users greater control and customization over the chatbots they create, allowing the creation of more complex bots with specific functionality. The Power Virtual Agents web editor, on the other hand, is designed to simplify chatbot creation with a more intuitive and user-friendly interface, making it easier for users to build quick and basic chatbots.

  • Multi-language support: The Bot Framework Composer supports multiple languages, allowing the creation of multilingual bots. In contrast, the Power Virtual Agents web editor focuses on a more limited selection of languages.

  • Licensing and Pricing: Bot Framework Composer is an open source tool that is free to use, while Power Virtual Agents Web Editor is a paid tool, requiring a subscription to Microsoft Power Platform or Dynamics 365.

In short, the Bot Framework Composer is a more complex and versatile tool that gives users more control and customization of their chatbots, while the Power Virtual Agents web editor is a more user-friendly tool focused on integration with Microsoft services. Both tools are valuable for building chatbots, but users need to choose the right tool based on their specific needs and requirements.


Power Virtual Agents is an effective tool for creating chatbots that can improve the user experience and automate common tasks. The addition of the GPT model in Power Virtual Agents allows chatbots to better understand natural language and generate consistent and accurate responses.

Additionally, the ability to embed chatbots into websites and apps means that chatbots can be a valuable tool for engaging with customers and improving the user experience. While the preview version of Power Virtual Agents with GPT has certain limitations and conditions, this tool has great potential to improve efficiency and user experience in a variety of business applications.

In future posts we will see the Bots in full action and we will dare to create one or the other!! Do not miss it!


bottom of page