What is the "Agentic Studio" Component in wolkvox Studio and How Does It Work
Table of Contents
Introduction
"Agentic Studio" is a component in wolkvox Studio that allows you to design IVR flows using natural language instructions. Instead of manually building the entire logic from scratch, the user converses with an artificial intelligence that interprets the functional requirement and generates the flow structure, the PHP code, and a readable version of the logic for review.
Its purpose is to accelerate the creation of IVRs, reduce syntax errors, and support the development of more complex flows, including menu scenarios, validations, transfers, and evolution from traditional tone-based models (DTMF) to more advanced experiences with ASR and TTS.
Currently, this component is only available for voice (IVR) routing points and is located within the "Cognitive" components group.
What is Agentic Studio?
Agentic Studio functions as a technical copilot within wolkvox Studio. Its main value is that it translates a business need into functional logic without the designer having to manually build the entire flow from the beginning.
With this component, you can describe, for example:
- The options in a menu.
- Wait times.
- What should happen if the user does not input anything.
- What data should be requested.
- How to confirm an input.
- Which skill to transfer the call to at the end.
Based on this conversation, the AI proposes the flow structure and asks for additional definitions when it detects missing important information, such as system voice, validations, confirmations, or transfer destinations.
What Is It For?
Agentic Studio is used to:
- Speed up IVR creation, reducing technical development time;
- Automate the writing of complex logic in PHP;
- Reduce critical errors, especially syntax errors and session variable handling;
- Facilitate the construction of menus, confirmations, and transfers;
- Support the migration from DTMF IVRs to more conversational experiences with ASR and TTS;
- Preserve key data logic from the start, such as identifications, contact variables, or service routes.
In practical terms, it helps convert a functional idea into a flow base ready for review and adjustment.
Configuration
- Open the component's configuration panel by double-clicking on it.
- There, you will find two main areas:
- A chat history section, where the conversation between the user and the AI is displayed;
- A field to write the prompt or initial instruction.
- There is also a "Code" tab, where the generated result is presented.

Write Your Requirement in Natural Language
In the "Design Your Flow (Write What You Imagine)" prompt field, you should describe what you want the IVR to do. Each prompt is sent using the "ENTER" key. You can formulate instructions such as:
- What options the menu will have.
- What should happen if the user does not respond.
- What data should be requested.
- How to confirm an entry.
- If it should be transferred to a skill.
- If the call should be ended.
The clearer the requirement, the better the AI can respond.

Use the Microphone If You Prefer to Dictate
The panel includes a microphone icon button that allows you to transcribe speech to text to dictate the prompt instead of typing it manually.

- You have a timer and timeline with audio waves available to check that the voice is being captured.
- Once you finish, click the "Send" button.
- If you want to restart the recording, click the trash can icon button.

Continue the Conversation with the AI
The interaction is not limited to a single message. After your initial instruction, the AI may ask additional questions to complete the flow logic. For example, it may ask you to define:
- Which catalog voice you want to use,
- What should happen after a selection,
- If an identification number should be captured,
- If data confirmation is needed,
- If the call should be transferred to a specific skill (agent queue).
This means the flow is built as a guided conversation, not as a closed configuration from the first message.


Review the Generated Result
- When the AI finishes building the proposal, you can review the result in the "Code" tab. There, you will see the generated logic in PHP. In addition to the code, the component is designed to provide you with a comprehensible structure that allows you to validate the flow's behavior before using it.
- When you finish the interaction and have reviewed the result, click "Save and Close."

What Types of Flows Can It Help You Create?
Agentic Studio can help you generate flows such as:
- Tone-based option menus;
- Data capture, such as identification number;
- Confirmation of entered information;
- Validation or error messages;
- Automatic closures due to wait time;
- Transfers to skills or specialists;
- Combinations of deterministic logic with AI, ASR, or TTS.
A typical example is an IVR where the user selects the type of document, enters their identification, confirms the number, and is then transferred to a specific area.
Usage Recommendations
For best results, it is advisable to:
- Write clear and concise prompts;
- Define the flow's intention from the beginning;
- Indicate what should happen in cases of error, timeout, or lack of input;
- Specify the desired voice if the flow will use TTS;
- Indicate which skill or destination the call should be transferred to, if applicable.
Important Considerations
Logic Validation
Although Agentic Studio speeds up flow generation, it is mandatory to review the generated logic before finalizing it. In particular, the pseudocode or expected behavior of the flow should be validated to confirm that it exactly meets the business need.
Additional Consumption
If the generated flow incorporates AI, ASR, or TTS components, this may result in additional consumption of credits or tokens, according to the current pricing.
Current Scope of the Component
Agentic Studio is currently only available for voice IVRs. It cannot be used, for now, in chat routing points or other types of flows.