Power Virtual Agent: Deployment and Analytics.

As we know, PVA is a new member in the power platform helps to easily create powerful bots using a guided, no-code, graphical interface, without the need for data scientists or developers.

In my earlier session, I talked about creation of BOT in Power virtual agent. We also discussed about how this BOT can be integrated with Power Automate. In this session we will discuss the details of BOT deployment. We will have a demo of integrating BOT with Microsoft Teams. Also, we will get in to the details of cool and exhaustive set of analytics to showcase the KPI's or the key performance indicators of the created BOT.

Deployment:

PVA provides an option to verify the functionality of the BOT on a Demo website. Once the BOT is ready then it can be deployed to the on multiple platforms or channels. There are 14 channels available where Power virtual Agent BOT can be published.

Analytics:

Analytics in PVA are categorized under three tabs which provides variety of charts and graphical views.

1. Summary

2. CustomerSatisfaction

3. Sessions

There is a limitation to the data available in the comprehensive set of analytics, details of last 30 days is available in the Key performance indicators (may change in future releases).

Also, sessions data can be downloaded at a maximum of seven days. If your bot had a high number of sessions, they will be broken down into multiple rows, each containing 2500 sessions.

Things to remember:

• BOT should be published atleast once, only then it can connect to a channel.

• Every time a change is made toa BOT, it should be published again in the PVA app and this willupdate the BOT across all the channels

• Different channel has differentend-user experience. So, Channel experiences should be consideredwhen a bot contents are authored.

For example:

Customer satisfaction survey appears as below:

To Summarize, Power virtual agent not only offers easy development but also easy configurable deployments of these BOT. Using Analytics, performance can be tracked to a granular level and this helps in modifying and enhancing the user experience which results in overall improvement.