3 metrics to monitor the performance of your bot on a Real-time dashboard

social crm

1577 views

3 metrics to monitor the performance of your bot on a Real-time dashboard

With the current rise of bots, companies are now increasingly investing in customer support via chat (WhatsApp or Messenger, for example). But how can I monitor customers that are using my brand’s bot?

 

Thinking about this, we integrated bots to Buzzmonitor Dashboards. This makes it much easier to analyze the performance of automated customer care, identify cases where the bot did not understand the questions of the user, and find out how many people are served by the bot.

 

What are bots?

Bot, short for robot, is a software created to simulate human interactions, they interact with users and allowing people to ask questions and providing appropriate responses.

 

1.Commands not understood: when my bot did not understand what the user meant.

Despite using scripts that facilitate interaction with users, bots still have some limitations, especially when the person you are interacting with it, makes a spelling mistake, or to request information that was not previously integrated into the repertoire of bot responses.

 

 

Thus, a report showing how many cases of "commands not understood" occurred during the week or month, it is easier to refine responses and feed your bot scripts with more information and response options.

 

2."Help" menu or the 9 key

The command “help” is the same thing as the 9 key when the consumer is being contacted using a telephone, and can be operated at any time to access the help menu or to speak to a human customer care consultant. Create a dashboard to monitor all people who interact with your bot and mention the word "help". Thus, being a case requiring greater complexity from the customer (once the bot functions as a mechanism of primary care for the solution of basic issues), you can identify users who need your help or have statistics showing you a search for the term help.

 


 

The only thing necessary is to classify the bot's responses with the tag "help".

You can also create an automatic alert through the Triggers function, so that each time a user sends the word "help" or any other important word for your bot, a notification is sent to the email of your analyst so he can follow up through a human customer care consultant.

 

 

3. Unique Users attended.

With a report showing the number of unique users who requested assistance from your bot, you can have the exact amount of customer care interactions attended.

 

 

This graph is an optimal indicator of the efficiency of your bot.


 

How to do it?

To create these reports all you need to do is to go to the Edit Menu option inside a Report. In the Variables tab,  on the “Facebook pages” option, select the Private Messages field within the account that has an active bot.

 

 

Talk to our E.life buzzmonitor team to integrate a bot to your Buzzmonitor.

 

TRY FREE BUZZMONITOR NOW

Comments

leave a comment