Agent Studio in HighLevel

Updated: February 4, 2026  | 
Author: Bill Stilwell

This post covers what Agent Studio is, how it relates to the individual AI agent types, what the workspace contains, and how agencies use it to deploy and manage agents across multiple clients efficiently.

Reading time: about 8 minutes.

What Is Agent Studio in HighLevel?

Agent Studio is HighLevel’s centralized workspace for building and managing all AI agents in a sub-account.

Before Agent Studio, different AI agent types were configured in different sections – Voice AI was in Settings, Conversation AI was in another part of Settings, and the two had no shared interface. Agent Studio consolidates that into one workspace where all agent types live, are created, and are managed together.

The result is a single place to see every AI agent in the account, what each one does, where it is deployed, and whether it is active – without hunting through multiple settings screens.

Find it in the left navigation of any HighLevel sub-account under Agent Studio or AI Agents.

What the Workspace Contains

The Agent Studio overview is an agent roster. Every configured agent in the sub-account is listed here – its name, type, assigned channels or phone numbers, and activation status.

From the overview you can create new agents, open existing agent configurations for editing, activate or deactivate agents, duplicate agents, and access the testing tools for each agent type.

The workspace is organized by agent type – voice agents and conversation agents are separated so the overview remains readable as the number of agents grows. This matters more for agencies managing clients with multiple deployed agents than for single-business users with one of each type.

The Three Agent Types

Agent Studio supports three agent types, each serving a different communication channel and use case.

Voice Agents handle phone calls – inbound and outbound – using AI-generated voice. Each voice agent has its own voice selection, call instructions, knowledge base, booking calendar, and transfer rules.

Multiple voice agents can be created within the same sub-account – for example, one for inbound lead qualification and a separate one for outbound reactivation calls.

Conversation Agents handle text-based conversations across SMS, live chat, Facebook Messenger, Instagram DM, WhatsApp, and Google Business Messages. Each conversation agent has a name, persona, objective, web crawl training, custom Q&A responses, channel assignments, and operating hours.

Flow-Based Agents are conversation agents enhanced with structured flows from the Flow Builder. When specific intents are detected, these agents follow defined branching paths rather than generating freeform responses.

They are built on top of the conversation agent system and share the same channel infrastructure.

Creating and Configuring Agents

Creating an agent in Agent Studio starts with selecting the type – Voice or Conversation – and giving it a clear internal name.

From there, the configuration interface for that agent type opens. All the fields covered in the individual agent posts – persona, objective, knowledge base, channel assignments, operating hours, transfer or handoff rules – are configured in the agent’s dedicated configuration screen, accessed directly from the Agent Studio workspace.

Agent Studio does not change the configuration options for each agent type. It changes where you find and access those configurations – through one workspace rather than scattered across multiple settings sections.

Testing Before Going Live

Agent Studio includes testing tools for both agent types.

For conversation agents, the Bot Trial interface lets you simulate text-based conversations with the agent before activating it on any live channel. Send test messages and evaluate whether the agent responds accurately, stays on topic, and maintains the right tone.

For voice agents, you can make a test call to verify the agent’s greeting, flow behavior, booking functionality, and transfer triggers. The call recording and transcript are logged to a test contact record so you can review the full interaction.

Testing before activation is not optional – it is the only way to know how the agent actually behaves when a real contact initiates a conversation or call. What looks correct in the configuration interface can behave differently in a live interaction.

Assigning Agents to Channels

Channel and phone number assignment happens from within Agent Studio.

For conversation agents, you enable each channel – SMS, Live Chat, Facebook Messenger, Instagram DM, WhatsApp, GMB – directly in the agent’s configuration. Each channel can be set to Auto-Pilot or Suggestive mode independently.

For voice agents, the assignment connects the agent to a specific phone number in the sub-account. When that number receives an inbound call, the assigned voice agent handles it.

For outbound calls, the agent is selected in the workflow Voice AI Call action.

The Agent Studio overview shows current assignments for all agents – so you can confirm at a glance which agents are active on which channels and numbers without opening each agent individually.

Duplicating Agents for Efficiency

The duplicate function in Agent Studio copies an existing agent’s configuration as a starting point for a new agent.

This is most useful in two scenarios. The first is when you need a similar agent for a different purpose in the same sub-account – for example, duplicating an inbound voice agent to create an outbound version with a different script and different objectives.

The second is when you are an agency setting up similar agents across multiple client sub-accounts. Build a well-configured template agent in one sub-account, note the configuration details, and recreate it efficiently in each new client’s sub-account using the same structure as a guide.

What Can You Do With It?

• See all your AI agents in one place: The Agent Studio overview shows every deployed agent – voice and conversation – their channels, and their status. No more checking multiple settings sections to understand what is active.
• Build multiple specialized agents for different purposes: Create a separate voice agent for inbound lead handling and another for outbound reactivation calls. Create a conversation agent for SMS and a separate configuration for live chat. Each serves its specific use case with the right configuration.
• Test before deploying to live channels: Use Bot Trial for conversation agents and test calls for voice agents before any real contacts interact with them. Catch configuration problems in testing, not in live conversations.
• Assign and reassign channels without hunting through settings: Activate, deactivate, or reassign channels directly from Agent Studio without navigating to separate phone number settings or channel integration settings.
• Duplicate agents as templates for new configurations: Instead of building each new agent from scratch, duplicate a well-configured existing agent and customize the specific fields that need to change.
• Manage AI deployments across a growing stack: As the number of deployed agents grows – multiple voice agents, multiple conversation agents, flows – Agent Studio keeps the full deployment visible and manageable from one location.

Key Definitions

Use Cases by Industry

Marketing Agencies – Multi-Agent Deployment

An agency sets up AI agents for 10 different client sub-accounts. Each sub-account needs a conversation agent for SMS and live chat, and some also need a voice agent for phone handling.

Agent Studio gives the agency’s team a clear view into each sub-account’s agent deployment – which agents are active, which channels they cover, and which need updating. Managing 20 agents across 10 clients from scattered settings sections would be operationally difficult.

From Agent Studio’s overview, it is manageable in a single session.

Result: The agency maintains visibility and control over a growing AI deployment stack without losing track of what is configured where.

Home Services – Specialized Agents by Call Type

A home services company runs two voice agents from the same sub-account. The inbound agent – configured for after-hours calls – qualifies service requests and books jobs.

The outbound agent – configured for reactivation campaigns – calls past customers with a seasonal offer.

Agent Studio shows both agents, their phone number assignments, and their activation status on one screen. When the seasonal campaign ends, the outbound agent is deactivated from the overview in one click – without affecting the inbound agent at all.

Result: Multiple agent configurations for different business purposes are managed cleanly without risk of interfering with each other.

Dental Practices – Channel-Specific Agents

A dental practice uses a conversation agent on Auto-Pilot for SMS lead capture and a separate conversation agent configuration in Suggestive mode for Facebook Messenger – where the practice manager prefers to review every AI-drafted response before it goes out.

Both agents are visible in Agent Studio with their respective channel assignments and modes. When the practice manager wants to switch the Facebook agent to Auto-Pilot for a week-long campaign, the change is made in Agent Studio without touching the SMS agent’s configuration.

Result: Different channel configurations can be adjusted independently without navigating through multiple settings menus or risking unintended changes to other agents.

Real Estate – Lead Source Segmentation

A real estate team has two conversation agents – one trained on buyer-focused content for contacts coming from buyer lead campaigns, and one trained on seller-focused content for contacts coming from seller lead campaigns.

Different phone numbers are assigned to each campaign source. Each number routes to a different conversation agent in Agent Studio.

Agent Studio’s overview confirms at a glance that each lead source is connected to the right agent.

Result: Lead source segmentation extends to the AI agent level – buyers get a buyer-trained agent and sellers get a seller-trained agent, producing more relevant first-contact conversations from each source.

SaaS Companies – Trial and Customer Onboarding Agents

A SaaS company uses HighLevel internally and deploys two conversation agents through Agent Studio: one for new trial signups focused on onboarding questions and feature discovery, and one for existing customers focused on support topics and renewal conversations.

Each agent has a different knowledge base, different persona, and different handoff rules. Agent Studio keeps both configurations visible and accessible without the team having to remember where each set of settings lives.

Result: Different customer lifecycle stages are served by appropriately configured agents – and maintaining both configurations over time is manageable from one interface.

Who Is This For?

How to Use Agent Studio

Step 1: Open Agent Studio

Navigate to Agent Studio or AI Agents in the left navigation of the sub-account.

The overview loads showing all currently configured agents with their types, channel assignments, and activation status.

Step 2: Review existing agents

Before creating anything new, scan the existing agent list. Confirm which agents are active, what channels they cover, and whether any need updating.

This prevents duplicating effort or accidentally creating a new agent that conflicts with one already deployed.

Step 3: Create a new agent

Click Create Agent. Select the agent type – Voice or Conversation.

Give it a descriptive internal name that identifies its purpose – for example, “Inbound Lead Voice Agent” or “SMS Conversation Agent.”

Step 4: Configure the agent

Complete the full agent configuration – persona, objective, knowledge base, channel or phone number, operating hours, and handoff rules.

Refer to the individual agent type posts for detailed configuration guidance. Agent Studio provides access to the same configuration fields covered in those posts, just from a unified workspace.

Step 5: Test the agent

For conversation agents, use Bot Trial to simulate text conversations. For voice agents, make a test call to verify the full call flow.

Complete the test before activating. Identify any gaps in training, persona, or flow and correct them before real contacts interact with the agent.

Step 6: Assign to channels or phone numbers

From within the agent configuration in Agent Studio, assign the agent to its channels or phone number.

For conversation agents, enable each channel and set the operating mode. For voice agents, assign to the specific phone number that should use this agent for inbound calls.

Step 7: Activate the agent

Toggle the agent to active. The agent immediately begins handling conversations or calls on its assigned channels and numbers.

Step 8: Monitor and refine

Check the Conversations inbox and call logs for the first batch of real interactions after activation.

Review how the agent handled actual contacts. Identify gaps and update the knowledge base, persona, or flow from the Agent Studio configuration screen.

Step 9: Duplicate for new use cases or clients

When a similar agent is needed for a different purpose or a new client, use the Duplicate function in Agent Studio.

Customize the name, persona, knowledge base, and channel assignments for the new use case rather than building the entire configuration from scratch.

How Does It Connect to HighLevel?

• AI Voice Agents: Agent Studio is where AI Voice Agents are created, configured, and managed. The Voice Agent configuration – voice selection, instructions, calendar, transfer rules – is accessed through Agent Studio.
• AI Conversation Agent: The AI Conversation Agent configuration – persona, web crawl, custom Q&As, channel assignments – is also managed through Agent Studio. The workspace consolidates both agent types.
• Conversation AI Flow Builder: Flows built in the Conversation AI Flow Builder connect to conversation agents managed in Agent Studio. Flow-based agents are conversation agents with structured paths layered on top.
• Workflow Builder: The Voice AI Call action in Workflow Builder references voice agents created in Agent Studio. When a workflow reaches the Voice AI Call action, it uses the agent configured in Agent Studio to handle the outbound call.
• Conversation AI: Agent Studio is the consolidated interface for the capabilities covered in the broader Conversation AI feature – bringing agent creation, training, testing, and channel management together in one workspace.

Common Questions

What is Agent Studio in HighLevel?

A centralized workspace for building and managing all AI agents in a sub-account – Voice Agents, Conversation Agents, and flow-based agents – in one interface. Replaces navigating to separate settings sections to configure each agent type.

Where is Agent Studio in HighLevel?

In the left navigation of the sub-account under Agent Studio or AI Agents. Some account versions may show it under Settings depending on the navigation layout.

What types of agents can I build in Agent Studio?

Voice Agents for phone call handling, Conversation Agents for text channel automation, and flow-based agents that combine conversation agents with structured Flow Builder paths.

Is Agent Studio different from Conversation AI settings?

Agent Studio is HighLevel’s newer unified interface that consolidates what was previously in separate Voice AI and Conversation AI settings sections. Depending on your account version, you may see Agent Studio as the primary interface or the older separate sections.

Can I create multiple agents in Agent Studio?

Yes. Create as many agents as needed within a sub-account – multiple voice agents for different call types, multiple conversation agents for different channels or purposes – each configured independently.

Can I duplicate or copy agents in Agent Studio?

Yes. Use the Duplicate function to copy an existing agent’s configuration as a starting point for a new one.

Customize the name, persona, knowledge base, and assignments without rebuilding everything from scratch.

Does Agent Studio include testing tools?

Yes. Bot Trial lets you simulate text conversations with conversation agents. Test calls let you call voice agents directly to verify behavior before assigning to a live number.

Can I assign an agent from Agent Studio to a phone number or channel?

Yes. Channel and phone number assignments are made from within the agent configuration in Agent Studio – no need to navigate to separate phone number or channel integration settings.

Does Agent Studio show agent performance data?

Agent Studio shows each agent’s status, assigned channels, and activation state. Detailed conversation logs and call recordings are accessed from individual agent records or the contact’s conversation history in the Conversations inbox.

Is Agent Studio available on all HighLevel plans?

Agent Studio is part of HighLevel’s AI feature set. Availability depends on account version and plan level.

Check the sub-account navigation for Agent Studio or AI Agents to confirm access in your account.

To Wrap It Up

Agent Studio is an organizational and workflow improvement as much as a feature in its own right.

The underlying capabilities – voice agents, conversation agents, flow-based agents – exist with or without Agent Studio. What the workspace adds is clarity and efficiency.

You can see everything deployed, everything active, and everything needing attention in one place rather than reconstructing that picture by navigating through multiple settings sections.

For single-business users with one conversation agent and one voice agent, Agent Studio is a minor convenience. For agencies deploying multiple agents across multiple clients, or for businesses with several specialized agents for different call types and channel configurations, it becomes genuinely important for operational control.

The duplicate function is worth noting specifically for agencies. A well-configured home services conversation agent can be duplicated and customized for a new home services client in a fraction of the time it would take to build from scratch.

The persona, knowledge base structure, and channel setup all carry over – only the business-specific content needs updating.

Here is how to get started:

1. Navigate to Agent Studio or AI Agents in your sub-account’s left navigation
2. Review the current agent roster – confirm what is active and what channels are covered
3. If you have no agents yet, create your first one by selecting the agent type and completing the configuration
4. Test before activating – Bot Trial for conversation agents, test call for voice agents
5. Assign to channels and activate
6. Monitor the first batch of real interactions from the Conversations inbox
7. Return to Agent Studio to make adjustments to the configuration based on what you observe
8. Use Duplicate when creating a similar agent for a new purpose or client

The overview screen is the most practical part of Agent Studio for ongoing management. Check it weekly when managing multiple agents – it is much faster than opening each agent’s configuration separately just to confirm what is active and where.