Introduction

The “Intents” component in wolkvox Studio allows you to route the customer through one path or another based on the result contained in a variable. In other words: here you define a set of intents (categories) and, for each one, its keywords or phrases. When the flow evaluates the content of the variable (for example, what the customer wrote in a chat or said in a call), the component selects the corresponding intent and continues through the associated branch.

This component is located in the Cognitive group and is available in all routing types: Voice, Interaction, Chat, and CRM + Webhook.

Configuration

1. Double-left-click on the component to open its configuration panel.
2. Configure the fields:
   • In the “Variable to analyze” field, type the variable that contains the text/result to be evaluated.
     • $txt_query: typically used to analyze the customer's written response in chat channels (WhatsApp, web chat, etc.).
     • $asr: used to analyze the customer's last response in a call, usually after using an ASR (speech recognition) component.
     • Recommendation: verify that the variable is actually receiving a value before reaching Intents (for example, by reviewing the flow and the order of components).
   • In “Intent”, type the name of the intent (this name will be what you see as the output branch of the component).
     • Best practice: keep intent names clear and consistent, as these names are usually reflected directly in the diagram paths.
   • In “Keywords or phrases (separated by commas)”, add the terms the system must recognize to assign that intent.
   • Click “Add Intent” to save it in the table.
     • Repeat the process for each intent you need.
   • At the bottom of the panel, you will find the “Activate NLP - AI” checkbox:
     • Without AI: you need to be more exhaustive with keywords/phrases to cover variations.
     • With AI: the component can interpret more broadly and associate text with an intent with greater flexibility, reducing the need to list many keywords.
     • Recommendation: even with NLP - AI, include representative keywords per intent to guide the behavior and reduce ambiguities.
   • Click “Save configuration” to apply the changes.

In the table where the intents are listed, right-click on an intent to:

• Edit Intent: change its name or adjust/add keywords and phrases.
• Delete Intent: remove it from the list.

Once you save, the component will be ready to route through its outputs:

Connect each intent to its next component (messages, validations, transfer, queries, etc.).

Make sure to preserve a path for “NO-MATCH” (when the system fails to associate the evaluated content with any configured intent).

Recommendation: for NO-MATCH, it usually works well to ask the customer for clarification, retry with a more specific guide, or escalate to another flow/service according to your strategy.