Why Microsoft Power Virtual Agent(PVA)?

Power Virtual Agent

Microsoft Power Virtual Agent is built on the Power platform and Bot framework. As we all know Microsoft Power platform is the low-code platform that spans Office 365, Azure, Dynamics 365, and standalone applications. Its innovative and valuable in almost every use case.

Power Virtual Agent has got the power from power platform and lot of features from Microsoft’s existing assets like Azure cognitive services and bot framework which enables PVA to extend the functionality for any complex scenarios. Most importantly all these assets are already matured, hence even PVA has a strong base on which it is built.

PVA has below advantages while working on chatbot implementation

Rapid chatbot development with no code : As its based on low code platform, for any simple conversation, most of the topics can be configured and no coding is required. Also for any actions connectors and Power automate can be used wherever appropriate. This accelerates the creation of the chatbot

SMEs and Developers can collaborate together to create the chatbot: Being no code solution, Microsoft does not leave the technical developers alone. Instead here SMEs and developers can work together on Topic developments. Configurations can be done by SMEs as they know the business and coding required for complex scenarios can be worked by technical developers

Extend the chatbot with Power automate and Azure bot framework for complex scenarios: Any actions can be implemented using Power automate flows with wide variety of connectors available which comes as no code implementation. However code full solution can also be developed such as developing custom skill using Bot framework

Availability of multiple channels for chatbot hosting and ability to integrate with live agent: We can develop single chatbot and host it on multiple channels such as Teams, custom websites, Facebook etc without spending much efforts. Even it can be integrated with human agent along with transferring entire conversation history for further reference

Built in analytics to analyze and enhance the functionality and usage: Every implementation can be enhanced by looking back at the history, analyzing the usage by out of box analytics in Power BI

Now we have understood why to use PVA. Lets see what are the considerations while implementing the PVA. Like any other implementations, we need to look at below pointers

Target Audience who are going to use the chatbot:

1. Small closed group within a department or Internal users where only Subject Matter Experts / citizen developers would do the configurations and it would be a rapid and easy development.

2. Internal Organization wide users like a support team and all employees within the organization. Here the IT team and citizen developers would be involved. The Bot would have SSO enabled and be available on either on intranet portal or teams etc. Entire ALM process need to be followed including development, testing and Go live. We may need to plan for actions to integrate with back end systems like ERPs, Office suits etc. This would be medium range of implementation

3. External users like customers for a manufacturing company. This would be a big project where developers , IT team and SMEs would be involved Here also we need to plan the ALM process in depth. Additionally the look and feel of the chatbot, scaling , performance are also important things to plan. API integrations, transfer to human agents would be essential to implement. Hosting channels should be appropriately identified. There might be complex scenarios where development using bot framework would be needed

Expectations on chat/conversations : We need to identify what is the source of these conversations, what the complexity of the conversations, any actions needed leading to integrations with APIs. is there need of escalation to human agent? all these points are to be considered

NFRs(Non Functional requirements): We need to check the NFRs like security, style, data retention, ALM, any need for custom analytics other than OOB

Development approach: Based on above parameters, we need to plan the development approach, however the proven way to follow in chatbot development is as mentioned below

With above information, we should be clear on our approach and scope of the chatbot implementation. Let’s move forward and go in depth of this implementation 😀