Home

How to configure the "Conversations" component in wolkvox Studio to create intelligent conversational flows

Updated at June 25th, 2025

+ More

Table of Contents

Symptom / Need Context / Scenarios Answer (Solution / Recommendation) Configuration for Voice-type routing rules AI Model Objective, tone of the conversation Configuration for ChatBot type routing rules Objective, tone of the conversation

Symptom / Need

In a world where customer service is increasingly automated, companies need tools that allow them to offer quick and personalized responses to their customers. However, setting up automated conversational flows can seem complex without an intuitive and efficient solution. How can a company implement dynamic flows that automatically respond to customer intent using artificial intelligence?

The answer lies in the "Conversations" component of Wolkvox Studio, a tool designed to create automatic conversational flows in both Voice (IVR) and Chatbot routing points.

Context / Scenarios

The "Conversations" component is ideal for scenarios where customer interactions require high personalization and precision. Some examples include:

Automated telephone support: A customer calls a veterinary clinic to schedule an appointment or request information about pet care.
Chatbot support: A user interacts with a chatbot on WhatsApp to inquire about available products or services.
AI-powered voice flows: A person communicates their request via voice, and the system automatically identifies their intent to direct them to the appropriate area.

This component allows you to design conversational flows that detect customer intentions and direct them toward specific paths, ensuring a fluid and efficient experience.

Answer (Solution / Recommendation)

The "Conversations" component is an advanced tool within Wolkvox Studio that allows you to create automated conversational flows using artificial intelligence. This component is especially useful for voice (IVR) and chatbot routing points.

Below is a breakdown of how this component can be used to improve customer interaction, depending on the routing point type:

Configuration for Voice-type routing rules

The Conversations component allows you to build automated conversational flows that combine artificial intelligence, speech recognition (ASR), and speech synthesis (TTS) to offer fully automated and personalized telephone support.

This component is key for scenarios where you want the customer to interact with the system using their voice naturally.

This component is ideal for:

Automate incoming calls with conversational responses.
Direct clients according to their spoken intention.
Implement intelligent voice bots for channels such as call centers, information lines, or self-service.

When using the "Conversations" component on Voice routing points, two additional options are enabled to enrich the interaction: TTS (Text-to-Speech) and ASR (Automatic Speech Recognition).

ASR (Automatic Speech Recognition)

Here you define the language the system will use for automatic speech recognition.

It's important to select the language appropriate for your target audience to ensure accurate and seamless detection of customer responses.
Proper ASR configuration ensures that the system understands customer requests, enabling effective and personalized interactions.
Spanish, Dutch, English, French, German, Hindi, Italian, Japanese, Korean, and Portuguese.

TTS (Text-to-Speech)

In this section, you can select the type of voice the system will use to "talk" to the customer.

You have a wide variety of voice options, including neural voices, which are more accurate and natural, improving the customer's listening experience.
The voice selected in this setting will be used throughout the conversational flow, providing clear and professional responses.
The appropriate selection of voice can be adapted to the brand identity or the preferences of the target audience, ensuring a more personalized and enjoyable interaction.

Interrupt TTS when detecting customer voice

TTS interruption allows the system to stop speaking as soon as it detects the customer has begun speaking. This simulates a more natural conversation. However, this feature has significant limitations.

This checkbox allows you to configure whether the system should interrupt text-to-speech (TTS) playback when it detects that the customer is speaking. Enabling or disabling it automatically depends on the stream structure.
- Disabling Interrupt Checking: Occurs if the component has a No-MATCH path.
- Interrupt check trigger: Occurs if the component has no No-MATCH path.
Recommendations:
- If you want to ensure a structured and secure response, use a "No-MATCH" route and allow the TTS to run completely before listening to the client.
- Only consider using the interrupt if:
  - You want a smoother conversation experience.
  - You are confident that the client will interact in quiet environments.
  - The stream is short and voice detection must occur quickly.
- Please note: Interrupts can be triggered by any sound, including background noise or people nearby. This can negatively impact the bot's understanding if it runs in noisy environments.

AI Model

The "AI Model" field allows you to select the type of artificial intelligence model that will be used to process customer interactions at the Voice routing point. This setting affects the speed and accuracy of intent analysis.

Available options:

Fast: This model prioritizes response speed, which is ideal for flows where faster interaction is required but may sacrifice some accuracy. Use the Fast model when minimizing response times and computational overhead is a priority.
Intelligent: This model prioritizes accuracy in intent analysis, which is useful for workflows where more detailed and accurate responses are required, although it may be slightly slower. Use the Intelligent model when a more precise analysis of customer intent is necessary, especially in cases where responses need to be highly personalized or complex.
Reasoner: Uses more advanced AI capable of processing complex reasoning and more elaborate language. Useful for sophisticated operations.
Automatic: The system analyzes the customer's input and automatically chooses whether to use the Fast, Smart, or Reasoning model as appropriate.

wvx Copilot Knowledge

The wvx Copilot Knowledge field allows you to select a domain trained in the wvx Copilot tool, so that the “Conversations” component can leverage that training as additional context.

You can select a pre-configured custom domain with instructions and documents in wvx Copilot.
If you choose the 'none' option, the system will not use any knowledge trained in wvx Copilot.

Objective, tone of the conversation

The "Objective, Tone of Conversation" tab will be available in both voice and chat routing points. Here you define the overall purpose of the chatbot, or the goal it should accomplish during the conversation. This includes the context or general type of response the bot should provide.

Important considerations

The prompt structure in the "Conversations" component is crucial to ensuring that the artificial intelligence understands and responds correctly to the customer's intentions. A well-designed prompt structure allows the AI to direct customers to the appropriate paths, improving efficiency and customer satisfaction. Below are the key considerations for building a good prompt structure.

Recommended structure for the prompt

Categories and Segments: Clearly define the categories and segments of intent the AI should handle. This helps organize potential queries and responses.
Keywords: For each intent, specify the keywords the AI should recognize. These keywords should be specific and relevant to the intent.
Intent Codes: Use unique codes for each intent, such as [CODVENTAS, [CODCITAS], [COD...], so that the AI can correctly identify and handle intents. This code format is mandatory to identify intents and should not contain special characters or spaces. It must fully respect the example structure [CODCATALOGO], [CODRECLAMO], [CODPRESTAMO], etc.
Messages and Responses: Provide clear and specific messages for the AI to use when it detects an intent. These messages must include the corresponding intent code.
Specific instructions: Include specific instructions for handling situations where the customer's query isn't listed in the FAQs. This ensures the AI can route the conversation to the correct intent.

Example of prompt structure

---

You are John Doe, the virtual assistant for wolkvox Pets. You represent the veterinary clinic's customer service team. Your mission is to provide clear, reliable, and friendly information about our services and products. Always maintain a professional, patient, and respectful tone.

Category: 'wolkvox Pets' Veterinary Clinic

Segments:
1. APPOINTMENT SCHEDULING:
- Keywords: appointment, schedule, consultation, review.
- Action: If the client indicates an intention to schedule an appointment, you should first ask for their pet's name, the desired date, and the time.
- Once you have all this information confirmed with the client, say the following message and code: Give me a moment, I'll take you to the selected area [CODCITA].

2. VACCINATION:
- Keywords: vaccines, vaccination, vaccination schedule.
- Action: If the client mentions words related to vaccines, you should ask what type of vaccine they need and confirm the information.
- Once confirmed, say the following message and code: Give me a moment, I'll take you to the selected area [CODVACUNAS].

3. CARE AND FEEDING:
- Keywords: care, nutrition, health, advice.
- Action: If the client is looking for information about care or feeding, you should ask what type of food they currently use for their pet and confirm the information.
- Once confirmed, say the following message and code: Give me a moment, I'll take you to the selected area [CODALIMENTACION].

4. EMERGENCIES:
- Keywords: emergency, urgency, immediate help.
- Action: If the customer mentions an emergency, you should ask for brief details about the situation and confirm the severity.
- Say the following message and code: I understand this is an emergency. I will connect you with a veterinarian immediately [EMERGENCY CODE].

5. SPEAK TO AN AGENT:
- Keywords: talk to an agent, I want to talk to a human, I don't want to talk to a bot.
- Action: If the customer wishes to speak with an agent, say the following message and code: Please allow me a moment, I will connect you with one of our advisors to help you with your request [CODAGENTE].

6. NO-MATCH (Unrecognized Response):
- Action: If you don't understand the customer's response, ask for confirmation. Example: "I didn't understand your request. Could you please try again? For example, you could write 'I want to schedule an appointment' or 'I would like to speak to an agent.'"

Additional context:
- Always use a friendly and professional tone.
- Prioritize emergency-related inquiries over other topics.
- Do not provide medical information without prior confirmation.
- Use the available variables to tailor each response to the customer's specific information. Call the customer by their first, middle, or last name occasionally to ensure the customer feels the personalized attention. Example: "I'd be happy to help you, Daniel! Let's solve this together."

Intentions and variable extraction

The "Intents and Variable Extraction" tab will be available for both voice and chat routing points. Here you define intents to create the different routes that can be accessed from the "Conversations" component, their keywords to facilitate the identification of the customer's intent, and variables to store necessary information for future use.
The "+" button allows you to add new intents to the conversational flow. Each intent defines a specific purpose or request the client can make, and is configured with keywords, variables, and prompts to instruct the artificial intelligence (AI) on how to process and respond to that intent.
The table shows all the intents defined in the "Conversations" component. It has three main columns:
- "Intent" column: Contains the name of the intent, which must be enclosed in square brackets (e.g., [CODCITA]). This name is crucial for linking the intent to other parts of the flow and for generating appropriate responses.
- "Words" column: Lists the keywords, separated by commas, that the AI will use to identify the customer's intent. The words should be specific and relevant to avoid confusion between intents.
- "Variables" column: Displays the variables associated with each intent. These variables are used to store information extracted from the conversation based on the provided prompts.

Clicking the "+" button will open a new configuration window where you can define the details of the intent.

Field " Intent ":
- Description: The name of the intent is specified here. It must be enclosed in brackets and without any special characters or spaces (for example, [CODCITA]).
- Example: [CODENERGY]
- Important: The intent name must be unique and clear to facilitate its identification within the flow.
Field " Comma-separated words for intentions ":
- This field lists the keywords the AI will use to detect customer intent. Keywords should be separated by commas.
- Example: For the intention [CODALIMENTACION], you could write: care, nutrition, health, advice.
- Considerations:
  - Avoid duplicating words between different intents to ensure AI doesn't get confused.
  - Use specific and common keywords depending on the type of client.
"+" button: This allows you to add variables with their respective prompts so the AI can save certain information in each variable created within the intent. Below this button is the variables table, which displays the variables added to the intent along with the prompts for each.
Once you have the fields configured, click on the " Add intent " button.

After configuring your keywords, you can add variables to extract specific information from the conversation. This is done using the "+" button below the variables table.

Variable Name: Enter the name of the variable that will be used to store the extracted information here. Don't include the $ sign (the system will add it automatically).
- Example: pet_name
Prompt variable: This field contains the prompt that instructs the AI on what information to extract from the conversation and save in the corresponding variable.
- Example: Save the name of the pet mentioned by the customer in this variable.
- Note: The prompt should be specific and focused solely on data extraction. Do not include additional instructions such as responding to the client, as this could cause errors in execution.
Add Variable: Clicking this button adds the variable with its respective prompt to the list of variables associated with the intent.

Complete example of setting an intent

Intention: [CODENERGY]
- Name of the intention: [CODALIMENTACION]
- Keywords: nutrition, feed, diet, feeding, ration, dietary recommendations, feed, kibble, wet food.
- Variables:
  - Variable 1:
    - Name: food type
    - Variable Prompt: Detects and saves the food type the customer mentions during the conversation, such as 'dry', 'wet', 'raw', etc.
  - Variable 2:
    - Name: preferred brand
    - Prompt variable: Identifies and saves the food brand that the customer prefers or mentions.
  - Variable 3:
    - Name: special requirements
    - Variable Prompt: Save any special requirements the customer mentions, such as allergies, specific dietary needs, or preferences.
Intention: [CODCITAS]
- Name of the intention: [CODCITAS]
- Keywords: appointment, schedule, consultation, check-up, reservation, schedule, availability, date, agenda, veterinarian.
- Variables:
  - Variable 1:
    - Name: petname
    - Prompt variable: Stores the name of the pet for which the appointment is being scheduled.
  - Variable 2:
    - Name: fechacita
    - Prompt variable: Detects and saves the date the customer wants to schedule the appointment.
  - Variable 3:
    - Name: horacita
    - Variable Prompt: Identifies and saves the client's preferred appointment time.
  - Variable 4:
    - Name: little motive
    - Prompt variable: Saves the reason why the customer is scheduling the appointment, such as 'annual checkup', 'vaccination', 'health problem', etc.

In the "AI Response" field, you can find the variable that contains the entire response that the artificial intelligence provided to the client.

Don't forget to save your changes once you've fully configured the component.