LinkedIn
Copied!

Table of Contents

Augmented Agent configuration

Version:

Only available versions of this content are shown in the dropdown

Managers with admin credentials can perform all of the configuration activities for Augmented Agent, as well as CSR activities during a live chat session. CSRs can only perform activities that involve live chat interaction and work with the Interaction Portal.

For CSR-specific activities, see Augmented Agent during live chat

Enabling Augmented Agent components

There are two types of Augmented Agent components: Augmented Agent for suggested replies, and the Augmented Agent Assistant. The Augmented Agent for suggested replies fetches responses from common phrases, page push entries, and knowledge articles and discussions, and the Augmented Agent Assistant sends a form to customers to fill out during their live chat session, to collect necessary information.

Enable Customer Decision Hub option under Suggested actions to enable the suggested replies option in the Augmented agent in App Studio.

To configure the components of Augmented agent, perform the following steps:

  1. In the navigation pane of App Studio, click Settings Chat and messaging .

  2. Under Settings: Chat and messaging, click the Augmented agent tab to view the Augmented Agent settings.

  3. To enable the Augmented Agent Assistant component, so that you can send a form to a customer's chat window during a live chat interaction to collect data, select the Enable Assistant check box.

    When you enable the Augmented Agent Assistant, it becomes available for all service cases. To disable the Augmented Agent Assistant for a specific service case, or an individual case, see Disabling the Augmented Agent Assistant.

  4. To enable the Augmented Assistant component that suggests automated responses to the CSR, select the Enable suggested replies check box, and then configure which suggested reply features to support.

    Pega Customer Decision Hub is a separately licensed product that is required in order to use the Augmented Agent suggested replies feature. If you have not selected the Enable Customer Decision Hub check box on the Suggested actions page, then the Enable suggested replies check box in the Augmented agent page is disabled.

    When you select the Enable suggested replies check box, you will see the following options:

    • Common phrases
    • Page push
    • Knowledge articles/discussions
    • Suggest content to agents at
    • Allow auto-response
    1. To include common phrases in the replies that Augmented Agent suggests to the CSR, select the Common phrases check box.

      When the customer enters any of the following questions:
      • What happens to my bill when I pause my service?
      • What is my financial impact when I pause the service?
      • Is there any fee for pausing my service?
      The agent suggests the following response from Common phrases: You will be charged a nominal fee of $20 for pausing the service. Do you want to go ahead and pause the service?

      You can pre-train the common phrases against the customer utterances in the Interaction Portal. When a customer enters a question, the agent matches the text in the question to a pattern. The agent then suggests a relevant response to the CSR.

    2. To include page push entries in suggested replies, select the Page push check box.

      When you enable this option, the CSR can redirect customers to a pre-determined destination, such as a web page or a help site, during an interaction.
    3. To include Pega Knowledge articles in the suggested replies, select Knowledge articles/discussions check box.

    4. To set the confidence threshold score for the replies that the Augmented Agent suggests to CSRs, move the Suggest content to agents at slider.

      Customer Decision Hub gives a score to that Augmented Agent provides to the CSR, based on the relevance of the suggested reply to the customer utterances in the chat session. When the confidence score exceeds the threshold that you set, the agent suggests the response to the CSR. If the score is below the threshold value, then there is no suggestion.
    5. To make Customer Decision Hub send automatic suggestions as responses when the confidence threshold for an automatic response is reached, select the Allow auto-response check box.

      The Respond to end-users automatically at slider appears.
    6. To set the confidence threshold score that triggers the automatic responses, move the Respond to end-users automatically at slider.

      When you select the Allow auto-response check box, the agent automatically sends replies to the end-user as the accuracy of the suggested replies increases, and the confidence level exceeds the specified threshold.
      When using the Allow auto-response option, a CSR cannot edit or cancel a response.
  5. Click Save.

Configuring common phrases for Augmented Agent

To decrease the time needed for a customer service representative (CSR) to respond to common customer questions, and to ensure consistent responses, you can configure a set of common phrases that CSRs can select from and send in their replies.

  • Enable common phrases in Settings Chat and messaging Augmented Agent in App Studio.
  • Set the required analytics repository in Settings Prediction Studio settings in Prediction Studio.
    The repository defaultstore is the default analytics repository. Choosing defaultstore as the analytics repository can result in data loss because it points to a temporary directory. To avoid data loss, choose a dedicated repository.

To help CSRs to more quickly select the most appropriate common phrase, the application groups the phrases into user-defined categories. During an interaction, the CSR clicks the Phrases button, selects from the list of categories, and then selects a common phrase from within that category.

When a CSR is engaged in a live chat or messaging session with a customer, the Augmented Agent suggested replies feature suggests common phrases to the CSR, based on the customer's chat input. For example, if the customer asks for information about their account, the Augmented Agent can suggest a common phrase that includes the requested account information. In addition, the Augmented Agent can be configured to send the common phrase automatically if the confidence level of the response is high.

To set up common phrases more quickly, you can import and export individual common phrases or all of the common phrases that you define.

Configuring categories for common phrases

Configure the categories that your organization uses to group the common phrases. During an interaction, CSRs view phrases by category so that they can more quickly select the most appropriate phrase.

  1. Log in to your application as a manager.

  2. In the navigation pane of the Interaction Portal, click Application Tools.

    The Common phrases tab opens by default.
  3. Click Actions Edit categories , and then perform one of the following actions:

    • To add a category, click Add category, and then enter the name of the category in the New category field.
    • To modify a category name, click in the name field and change the name.
    • To delete a category, click the Delete icon next to the category name.
      Deleting a category deletes the common phrases in that category. When you delete a category, the application displays the number of common phrases that category contains so that you can move the phrases to a different category if necessary.
  4. Click Submit.

Configuring common phrases

Configure the common phrases that a customer service representative (CSR) can select from during chat, messaging, and email interactions.

To copy existing common phrases from another environment, or to copy individual common phrases, see Importing and exporting common phrases section below.

  1. Log in to your application as a manager.

  2. In the navigation pane of the Interaction Portal, click Application Tools.

    The Common phrases tab opens by default.
  3. Select whether to add a new common phrase or edit an existing common phrase:

    • To add a new common phrase, click Add phrase.
    • To edit an existing common phrase, click the Edit phrases icon next to the phrase.
  4. In the Phrase name field, enter a short name for the phrase that clearly identifies the phrase to the CSRs.

    During an interaction, the CSR sees only this short name and must determine the correct common phrase to select based on this short name.
  5. In the Phrase category list, select the category to which this common phrase applies.

    During an interaction, the common phrases are grouped by category so that the CSR can more quickly find the appropriate response.
  6. To restrict the selection of a common phrase to customers with specific browser locales, select the language of the phrase in the Language list.

    By default, phrases apply to all languages and locales.
  7. To restrict the selection of a common phrase to customers using a specific channel, select that channel in the Channel list.

    • To restrict the phrase to chat and messaging interactions only, select Chat.
    • To restrict the phrase to email interactions only, select Email.
    • To allow the phrase in all supported interaction types, select All.
  8. To exclude the phrase from Augmented Agent automatic responses, select the Exclude from auto response check box.

    Automatic responses are part of the Augmented Agent suggested replies feature.
  9. In the Phrase text field, enter the complete text that will be sent to the customer when the CSR selects this common phrase.

    To insert properties that will be populated at runtime, such as the CSR name or customer account information, complete the following steps:
    1. Click Insert attribute.

    2. Select the attribute category from the Refer to list.

    3. Select the property name from the Attribute list.

    4. Click Append.

      The Phrase text field shows a placeholder where the property will appear when the common phrase is sent.
  10. Optional:

    To train the Augmented Agent suggested replies feature to recognize when to suggest this common phrase to a CSR, complete the Seed customer utterances section as described in the Configuring customer utterances to train the Augmented Agent section below.

  11. Click Submit.

Pre-training common phrases with customer utterances

Training the suggested replies feature of Augmented Agent consists of defining the customer inputs, also known as customer utterances, that will trigger the Augmented Agent to suggest a response from a common phrase.

The following figure shows the Phrase name Pause service, and the Phrase category Billing in the Add phrase dialog box. You configure the customer utterance by clicking +Add utterance, and configuring the customer input, for example:

  • What happens to my bill when my service is paused?
  • What is my financial impact when I pause the service?
  • Is there any fee for pausing my service?
  • What is the cost of pausing my service?
Customer utterances and suggested reply
Customer utterances and suggested reply

You must also configure the Phrase text, which enables the Augmented Agent to suggest possible responses in chat interactions. For example, the agent suggests that Yes, there is a small service fee of $20 that you will continue to pay each billing period. Do you want to go ahead with the pause service? for any of the customer utterances listed above. Most of the time, the Augmented Agent takes the burden of typing the entire reply off the CSR. The CSR can choose to send the suggested reply, edit the reply, or create a new reply, if the suggested response is not completely appropriate in the context of the customer utterance.

The customer utterances that you define to match common phrases do not have to match the actual customer input to trigger a suggested reply. For example, if the customer input is Do I have to pay any fee for pausing my service?, the Augmented Agent matches the pattern of the customer input with the closest customer utterance, and then triggers the suggested reply even if the customer utterance that you define under Seed customer utterances differs from the actual customer input in the chat interaction.

Configuring customer utterances to train the Augmented Agent

During a live chat or messaging interaction with a customer, the Augmented Agent suggested replies feature can suggest common phrases to a customer service representative (CSR) based on the customer chat input. Training the Augmented Agent suggested replies feature consists of defining the customer inputs, also known as customer utterances, that will trigger the Augmented Agent to suggest a common phrase.

Complete the following procedure to train the Augmented Agent suggested replies feature to recognize specific customer input so that the Augmented Agent can suggest common phrases in response.

  1. Log in to your application as a manager.

  2. In the navigation pane of the Interaction Portal, click Application Tools.

  3. Open the common phrase that you want to train by clicking the Edit phrases icon next to the phrase.

  4. In the Seed customer utterances section, click Add utterance.

  5. Enter a phrase that, when entered by a customer during a live chat or messaging interaction, triggers the Augmented Agent suggested replies feature to suggest this common phrase to the CSR, or send it automatically if the automatic response feature is enabled.

  6. To import customer utterances for a phrase, do the following steps:

    1. In the Seed customer utterances section, click the Click to Import/Export icon. The system displays a pop-up window.

    2. In the pop-up window, select the spreadsheet that contains the customer utterances that you want to import.

    3. If you want to verify the format in which you can enter the customer utterances in a spreadsheet, click Download sample utterances file.

    4. Select either of the following import options:

      • Add to existing utterance: Adds the imported customer utterances to the existing utterances.
      • Replace existing utterances: Replaces the existing utterances with the ones in the imported file.
  7. In the Seed customer utterances section, click the Click to Import/Export icon, and then click Export to export customer utterances of a phrase.

    The application opens a browser and downloads the customer utterances data for a phrase as a spreadsheet. You can find this spreadsheet in your local Downloads folder.
  8. Click Submit.

Importing and exporting common phrases

To support the faster setup of common phrases, you can import and export individual common phrases or all common phrases. Importing or exporting is especially useful when copying common phrases from one environment to another. If you have already trained the Augmented Agent to suggest common phrases, this is a faster way of moving the training information from one environment to another.

  1. Use your manager credentials to log in to the application environment that contains the common phrases that you want to import.

  2. In the navigation pane of the Interaction Portal, click Application Tools.

  3. On the Common phrases tab, click Actions Export phrases to export common phrases.

    The application exports the data to a spreadsheet file in your local Downloads folder.
  4. On the Common phrases tab, click Actions Import phrases to import common phrases.

    Select the spreadsheet files that contains the common phrases that you want to import, and then click Open.
    The common phrase or phrases are added to the current environment.

Configuring Knowledge Management articles in Pega Customer Service Chat

Knowledge articles provide relevant content to the CSR while working on a case. These articles avoid the need to perform time-consuming searches and provide necessary information to the CSR.

To use this option, create knowledge articles in the Knowledge Portal. Import the KMSampleData.jar into Dev Studio to access the knowledge articles that you create. Configure the articles when you configure the web chatbot.
  1. To enable the knowledge articles/discussion option, see the Enabling Augmented Agent components section in Augmented Agent configuration.

  2. To select the help categories that appear during a live chat, see the Configuring Pega Knowledge help categories in Web Chatbot section in the Configuring the Web Chatbot.

  3. To configure the coaching tip and associate the tip with knowledge articles, see the Associating a coaching tip with a knowledge article section in Augmented Agent during live chat.

Have a question? Get answers now.

Visit the Collaboration Center to ask questions, engage in discussions, share ideas, and help others.