Symptom / Need

Businesses need a tool that allows them to intelligently analyze and respond to customer inquiries across various communication channels. The wvx Copilot component in Wolkvox Studio offers an advanced solution for managing these interactions using artificial intelligence.

Context / Scenarios

In today's business environment, customer service is crucial to maintaining user satisfaction and loyalty. Wolkvox Studio, as part of the contact center service configuration software, provides tools for designing customized interaction flows. The wvx Copilot component, formerly known as Mr. Wizard, allows you to analyze customer messages and provide automated responses based on prior training. This component is useful in various scenarios, such as:

• Chatbots: Interactions via WhatsApp, web, Facebook, Instagram, Telegram, and SMS.
• Voice (IVR): Transcription and analysis of voice messages.
• Interactions: Managing emails and interactions on social media such as Facebook, Instagram, and Twitter.
• CRM and Webhook Integrations: Automating action-based workflows in your CRM or external applications.

Answer / Solution

The wvx Copilot component in Wolkvox Studio is an advanced tool that allows you to improve customer interaction management through artificial intelligence. Here's how it works and how to configure it:

• Domain Selection:
  • 'Wolkvox Copilot Domain' field: Here, you select the wvx Copilot domain that has already been created and trained with PDF documents. This domain defines the context and knowledge the AI will use to analyze the messages.
• Question setup:
  • 'Ask wvx Copilot' field: This field defines the variable containing the customer message you want to analyze. The variable can be $txt_query for chatbots, a speech-to-text transcription variable for IVR (e.g., obtained through the ASR component), or any other variable containing the message.
• Tests and answers:
  • 'Test' button: Allows you to test the component's response by typing a question in the 'Ask wvx Copilot' field. The answer will appear in the 'Test Response' field.
  • 'Response Variable' field: Contains the name of the variable that will store the final wvx Copilot response, which can be used in other components to display it to the client.

Example of using the 'wvx Copilot' component within a flow in a ChatBot type routing point

The main objective of this flow is to provide an efficient and automated customer service experience using the wvx Copilot component to answer specific financing, general, and sales inquiries. The flow is designed for chatbot-type channels (WhatsApp, Web Chat, etc.), where customers can interact naturally and receive personalized responses based on their query.

The flow is structured step by step as follows:

1. Startup Component:
   1. The flow begins with the Start component, which is unique and mandatory for all routing points. This component marks the client's entry point into the flow.
2. 'Reply Chat' Component:
   1. The Reply Chat component is configured with the message: "Welcome to wolkvox Cars. Please write your question." This message is sent to the customer to initiate the interaction and prompt them to write their question.
3. 'Intentions' component:
   1. The next component is Intents, which is configured to detect three types of keyword-based queries:
      1. Financing: Keywords such as "finance," "credit," "loan," and "installments."
      2. General Question: Keywords such as "branches," "headquarters," "cities," "dealers," and "countries."
      3. Sales: Keywords such as "models," "cars," "vehicles," "motorcycles," "prices," and "buy."
   2. Depending on the detected intent, the flow is divided into four possible routes:
4. Intent-based routes:
   1. Route 1: NO-MATCH
      1. If no intent is detected, the flow follows the NO-MATCH route, which returns the client to the Reply Chat component to reformulate their query.
   2. Route 2: Financing
      1. If the financing intent is detected, the flow is directed to a wvx Copilot component configured with the financing domain. This component parses the customer's message (captured in the $txt_query variable) and generates a response, which is saved in the $res_MrWizard variable. A Chat Reply component then displays the response generated by wvx Copilot to the customer.
   3. Route 3: General Question
      1. If the general question intent is detected, the flow is directed to a wvx Copilot component configured with the default domain. This component parses the client's message (captured in the $txt_query variable) and generates a response, which is saved in the $res_MrWizard variable. A Chat Reply component then displays the response generated by wvx Copilot to the client.
   4. Route 4: Sales
      1. If the sales intent is detected, the flow is directed to a wvx Copilot component configured with the sales domain. This component parses the customer's message (captured in the $txt_query variable) and generates a response, which is saved in the $res_MrWizard variable. A Chat Reply component then displays the response generated by wvx Copilot to the customer.
5. End Component:
   1. All three paths (financing, general inquiries, and sales) end in an End component, which is mandatory in all flows. This component marks the end of the interaction.

Note: This is a very basic example, but it's ideal for understanding the possibilities of flows that include the wvx Copilot component. The possibilities are extensive depending on the purpose of the flow and the different options available to the client entering the flow. With wvx Studio, you can design more complex and customized flows to suit the specific needs of your business.