Skip to main content

Email routing overview

Introduction

In this video, you will learn how Pega Infinity™ uses AI-powered text analytics to do intelligent email routing.

Video

Transcript

In this video, you will learn how Pega Infinity™ uses AI-powered text analytics to do intelligent email routing.

Customer satisfaction is a reflection of what a customer expects from a company vs. what they experience from the company. Meeting, or even better, exceeding customer expectations means addressing their service requests and complaints promptly, accurately, and with personalized solutions. Doing this will ensure the customer has a great experience.

Assume a company has an email channel in which its customers can send any type of email—from service requests to compliments to sales inquiries—to the product support team. These emails are often routed to the same container and are uncategorized. This makes it difficult to provide each customer with a prompt and personalized response.

In Pega Infinity, this issue can be addressed by routing the emails using artificial intelligence. Pega Infinity uses AI-powered Natural Language Processing to detect the topic of an email and route the email to the appropriate container.

Consider emails from customers requesting an account address change, making a compliment, or requesting a new credit card. With the help of AI-powered text analytics, Pega Infinity is able to read and understand the content of each email and route it appropriately. This means customer service representatives can be alerted to any account-related service requests and resolve them quickly.

Email routing is done using the topic detection mechanism. The two types of topic detection are rule-based and model-based.

In rule-based topic detection the routing is based on the rules configured in the email channel. AI-powered text analytics is used to detect the topic of the email, and the channel rules route it to the right container. This type of topic detection may detect one or more topics if the email contains words associated with more than one topic.

Let’s consider a service request email from a customer. The email content is analyzed and routed to the right container. If an email from another customer contains words that are associated with two topics, the rule-based topic detection detects both topics. The email can then be routed to two different containers depending on how the channel rules are configured.

In model-based topic detection the routing is based on AI models built by a Data Scientist using machine learning. Building these models requires a training data set and a test data set. The data sets consist of a list of emails and the associated topic for each email. This type of topic detection identifies the most accurate topic based on the AI model and training set used by the Data Scientist.

Let’s consider the same service request email from the customer. The email’s content is analyzed and routed to the right container. If an email from another customer contains words associated with two topics during the training of the models, the model-based topic detection detects both topics but typically with a different accuracy factor. In this case, the topic with the highest accuracy factor is chosen.

Pega Infinity also enables you to extract entities from an email. This means that when an email is sent, certain entities such as account number, email address, street address, etc. can be automatically detected and extracted. This allows certain emails to be automatically processed or given priority.

Besides Topic detection and Entity Extraction, Pega Infinity uses its AI-powered text analytics to enable you to detect the sentiment of an email based on its content.

Suppose a customer sends the following email to customer service.

The email’s content is a combination of different types of sentences. The first sentence has a neutral sentiment. In the second sentence, the customer expresses his satisfaction with the connectivity, so it has a positive sentiment. The third sentence is negative, as the customer is complaining about an overcharge. The overall sentiment of the email is determined based on the weight of the content sentiments. In this specific example the overall sentiment is neutral with a sentiment score of 0.05.

The sentiment score is a value between -1 and 1. In the out-of-the-box configuration, a sentiment score <-0.25 results in a Negative sentiment, a sentiment score between -0.25 and 0.25 results in a Neutral sentiment, and a sentiment score above 0.25 results in a positive sentiment.

Once the email sentiment is detected, you can configure the email channel to route a specific topic with a specific sentiment to a specialized agent for a quick and personalized response. For example, you could route an address change with a neutral sentiment to a Service Agent, a complaint email with a negative sentiment to a Manager, and a credit card inquiry with a negative sentiment to a Financial Services Specialist.

 

In summary, Pega Infinity’s AI-based email routing capability enables customer service representatives to be more productive, reduces request processing time, and improves the customer experience by providing prompt and personalized service.

 

 

If you are having problems with your training, please review the Pega Academy Support FAQs.

Did you find this content helpful?

Want to help us improve this content?

We'd prefer it if you saw us at our best.

Pega Academy has detected you are using a browser which may prevent you from experiencing the site as intended. To improve your experience, please update your browser.

Close Deprecation Notice