Content of Article
Microsoft's answer to the increasingly popular chatbots is Power Virtual Agents. If you don't know what chatbots are, these are the small pop-up windows (usually at the bottom right) on a web page that offer help to the visitor of the web page. As the Power at the beginning of the name suggests, Power Virtual Agents is part of the Microsoft Power Platform. Therefore, a chatbot created with Power Virtual Agent can also be combined with Power Automate (formerly Flow) to trigger a workflow from a dialog.
The Microsoft Power Platform is a so-called "low-code platform". Such a platform allows business users to create software or features using a visual application designer, which usually requires development.
Power Virtual Agents maps the entire bot lifecycle. From creating the dialog to publishing in communication channels such as teams, slack, Facebook or Messenger. Let's take a look together at what the chatbot creation looks like when a potential customer wants to talk to a consultant.
First, you have to create a new topic. In our case, it is a request to make an appointment with a consultant. Then it is time to enter as many trigger sentences as possible on the topic. These are possible phrases that the potential customer writes in the chat window when he wants to speak to a consultant.
Once we have added some fitting triggers, we can create the actual dialog with its different branches. There should definitely be more than 7 triggers, as here in our example. If you want to play it safe with the triggers, send e-mail to your colleagues how they would formulate the question on this topic. If you try to create all trigger sets yourself, you end up with phrases that are all very similar.
The chatbot reacts to one of these triggers with a greeting and asks a question to limit the topic. Each possible answer creates another path that can take a different path.
The topic should be as narrow as possible so that you can name the right contact person to the visitor at the end of the dialogue.
Last but not least, we would like to know how high the available budget of the interested person is. Depending on how high the budget is, we would give the visitor different contact persons. It should be noted here that the figures are fictitious and we do not classify potential customers according to such criteria. For power agents, it is necessary to add a variable for such a case.
The possible variables are limited to 4 types:
We select the currency type for our case. With the type "Text", for example, it would be possible to ask the visitor for his name and then reuse it in an answer. This allows you to personalize the conversation a bit.
Our variables are as follows:
By the way, the order is important here. The highest number should always be at the beginning and then descending.
Depending on how large the budget is, the interested person receives another contact. If the budget is over 50,000, the interested person immediately receives the VP of Sales phone number. With less than 10,000, only a hint to write an e-mail.
This is what the result looks like in action:
Microsoft's Power Virtual Agents are ideal for typical customer service use cases. For more complex scenarios, such as a support case, the Builder is not necessarily suitable. If you want to connect the chatbot with an existing knowledge database, development is necessary and cannot be mapped with the existing Visual Builder. However, Microsoft has already announced that it will continue to expand Power Agents and make them attractive for use beyond customer service. By the way, Power Agents is based on the Microsoft Bot Framework. Should you reach your limits with the Visual Builder, the chatbot can be extended in the Bot Framework for more complex scenarios. Power Automate (formerly Flow), with its huge number of connectors to other applications and platforms, is particularly attractive for business users.