Edit

AI-based chatbot authoring overview

  • Article

Important

Power Virtual Agents capabilities and features are now part of Microsoft Copilot Studio following significant investments in generative AI and enhanced integrations across Microsoft Copilot.

Some articles and screenshots may refer to Power Virtual Agents while we update documentation and training content.

By leveraging recent advancements in AI large language models, Microsoft Copilot Studio transforms how you build chatbots, increasing a chatbot's usefulness and significantly reducing manual authoring and configuration. You can:

  • Create an instantly useful chatbot. With no manual authoring, an "empty" chatbot can respond to your users with relevant information, even if you haven't manually created a topic for it.

  • Author topics by describing what you want. Describe what you want your topic to do, and Microsoft Copilot Studio creates it for you, complete with conversational responses and multiple types of nodes. Use the suggested topic as is, or as a starting point for further iteration.

Prerequisites

  • You'll need an account for Microsoft Copilot Studio.

    Note

    If you don't have a Microsoft Copilot Studio account, or you haven't created chatbots with Microsoft Copilot Studio before, see the Quickstart guide for building bots with GPT.

  • You must be using the current version of Microsoft Copilot Studio, and the bot type mustn't be Classic. Classic chatbots have (classic) added to their name.

    Screenshot of the list of chatbots with one bot with clasic added to its name.

  • This capability may be subject to usage limits or capacity throttling.

Generate a bot's responses on-the-fly

Microsoft Copilot Studio uses an underlying natural language understanding model for understanding a bot user's question and determining the right topic. Typically, you create multiple topics within a bot to account for what a user might ask. But, sometimes, a bot user might ask a question for which there is no built topic.

When this happens, by default the bot prompts the user to rephrase their query. If, after two prompts, the bot still can't determine the user's intent, the bot escalates to a live agent through the system Escalate topic.

You can specify a system Fallback topic to customize the response and actions the bot takes - but this isn't always helpful for the bot user.

The Boost conversations option in Microsoft Copilot Studio helps to solve that issue by connecting the natural language processing capabilities of Azure OpenAI, which includes use of OpenAI's GPT technology, with the AI already in Microsoft Copilot Studio (including transformer-based natural language understanding) to:

  • Understand a user's intent by parsing what they type, and determining what they're asking
  • Find, collate, and parse relevant information from a URL you specify
  • Create a plain language response and deliver that to the bot user

In other words, if the bot doesn't already have an answer for what the user is asking, it can get the information and create one "on-the-fly".

See the Boost conversations topic for how to set up and use Boost conversations, along with tips for getting the best from the AI.

Create and edit topics by talking to AI

As part of creating a bot, a bot maker usually creates multiple, individual topics, with each topic corresponding to a particular issue or action.

This can require a large amount of time, and often requires specialized knowledge in the field of conversational user experiences, along with deep technical or domain knowledge. This isn't always possible in every organization - especially those that rely on fusion teams where multiple disciplines work directly in the Microsoft Copilot Studio authoring canvas.

The Create with Copilot option in the Microsoft Copilot Studio canvas lets bot makers simply describe what they want to achieve, and then produces a topic path that achieves that goal. It does this by combining the natural language understanding models already in Microsoft Copilot Studio with Azure OpenAI to:

  • Understand what the bot maker wants to achieve by parsing their request
  • Apply knowledge of how nodes within a topic work together, and how a topic should be constructed for the best effect
  • Generate a series of connected nodes that together form a full topic
  • Use plain language in any node that contains user-facing text that corresponds with the bot maker's request

This capability means that what previously might have taken several hours can instead take minutes to complete.

See the Create (and edit) with Copilot topic for how to set up and use the feature, along with tips to get the best from the AI.

Next steps