Setting up Pega Chat
Setting up Pega Chat
|Description||Suggestions for setting up Pega Chat to handle various customer interactions.|
|Version as of||8.4|
|Application||Pega Customer Service|
|Capability/Industry Area||Chat and Messaging|
Chatbot-assisted versus human-assisted customer service
Pega Intelligent Virtual Assistant allows the design and deployment of a chatbot that can be trained to assist your customers with their service requests. Pega chatbots offer great flexibility in substituting human agents for initiating conversations with customers, gathering their intent, and assisting with simple and conversational case processing, where relevant.
Pega chatbots employ natural language processing (NLP) to understand customer intent, but in some cases, for example, because of inadequate training of the bot, chatbots might be unable to process customer requirements reliably. So, it is important to distinguish between scenarios where a bot can be trusted to assist a customer and where grievances need to be escalated directly to human agents. At a minimum, bots can be trained to greet customers, provide basic business information, and even process service cases that involve simple inputs and outputs such as balance inquiry on a personal bank account.
It is recommended that you conduct an audit of all the different tasks that customer service representatives (CSRs) regularly perform when they engage with customers. Tasks that are repetitive, low on complexity, and high in volume generally can be assigned to well-trained chatbots. Communicating information about hours of operation, the status of pending service cases, and accepting requests to pause or restart a service are good examples of the kinds of tasks that a chatbot can perform. More complex assignments, such as troubleshooting a product defect or accepting a transaction dispute claim, can be directed to well-trained human agents, at least until your chatbots have undergone sufficient training and testing to inspire confidence.
Matching a customer issue to the right agent
Pega Chat operates primarily around a concept called queues. Queues within Pega usually serve as proxies to business functions such as billing. Service representatives with the required skills man the queues, ready to assist incoming customer inquiries of that particular nature. But queues need not always represent skills or business functions. You are free to utilize the pattern to organize and map customer requests with service resources to suit your business needs. For example, incoming chat requests could be separated into authenticated and unauthenticated requests and specifically processed to the identified customer context.
Customer intent can be gathered in the following ways:
- From an explicit request to select a queue
- Inferring the intent based on the webpage from where the chat is initiated, for example, a products page
- Processing available metadata such as the language of the calling browser or the channel from where the customer initiated the chat or even a selection made by an end-user (if you want to combine a queue indicating a question with language or channel source, for example)
Presenting customers with a list of available queues to select from is usually the most direct way to gather customer intent and subsequently map to the right agent.
The second feature is called URL mapping, which is useful when a customer needs to be directed to an agent with minimal perceived hindrance. Because this feature connects intent with a webpage, it is only available as an option in Web Chat and not in other Messaging channels. This configuration can be turned for a particular queue from App Studio > Settings > Chat and Messaging > Queues. For more information, see Configure chat and messaging queues.
The third feature gathers additional context through what we refer to as intelligent routing.
Managing the look and feel of the interface
It is important to utilize the available configurations to achieve the desired functionality instead of manipulating the source CSS and JS files. This helps with easy maintenance and upgradability of the application while delivering the desired look-and-feel functionality.
Handling customers during off-hours
For businesses that do not operate their contact centers 24x7, it is important to allow customers to register their issues offline and inform inquiring customers about standard hours of operation and, in some cases, downtime updates. Pega Chat offers easily configurable options to communicate such information with your customers, making for a good customer experience.
Pega Chat also allows you to offer your customers an option to send an email about their inquiry when your contact center is offline or when all the CSRs are too busy with other customers to respond quickly enough. Such customer-initiated offline interactions will be routed to an appropriate CSR when the center becomes operational again. It is useful to deploy this flexibility of using email as a channel when synchronous communication is not immediately feasible. Implementing the Pega Email functionality alongside chat is highly recommended given the flexibility it offers in handling customers even as the contact center is overwhelmed with incoming requests or constrained by CSR capacity.
Selecting the Allow end-customers to submit an email when chat is unavailable option from App Studio > Settings > Chat and Messaging > When Chat CSRs are not available initiates the offline email process setup. For more information, see Configuring the When chat CSRs are not available settings.
Follow-up messaging/Interaction ageing
Unlike with call-based interactions, CSRs will be able to assist multiple customers at the same time over text-based channels such as webchat. To control this function, Pega Chat allows concurrency configurations that can be applied at multiple levels such as Global, Queue, and CSR.
As agents toggle between multiple open text interactions, it is essential to keep track of all ongoing conversations to prompt customers and agents for action and even terminate interaction after prolonged periods of inactivity. This helps bring the focus of the customer and agent back to the ongoing interaction. Setting up a reasonable inactive time for termination of the interaction helps free up the agent's capacity to accept other customer inquiries that are queued for assistance. Because inactive chats quickly eat into the available CSR bandwidth and also increase the likelihood of vulnerabilities such as customer session hijacking, it is strongly recommended that you utilize these automatic controls to monitor and gracefully terminate customer interactions that are out of focus for too long.
These configuration options can be accessed from App Studio > Settings > Chat and Messaging > Follow up messaging. It is recommended to prompt users about inactivity around the 180-second mark and terminate the interaction after 300 seconds of persisted inactivity. For more information, see Configure follow up messaging.