What changed
Agent Studio has grown from a simple node editor into a more complete agent builder. Recent areas of development:
- More built-in node types, including conditional branches based on caller intent, contact data lookups and custom knowledge base queries.
- An in-app test interface that lets you simulate a conversation before going live. It saves time compared with placing a real test call.
- Better connection to the workflow builder. An agent can trigger a workflow at any point in the conversation, and workflows can invoke agent flows.
- Better logging and conversation replay, so you can see what the agent said and what triggered each decision.
This is a summary of the general direction. For exact build numbers and release dates, check the official release notes in your GoHighLevel account.
Who this affects
- Agencies building AI agent products. The test interface and better logging make it faster to build and debug an agent before handing it to a client. See the Agent Studio multi-step workflow build playbook.
- Businesses replacing phone trees. A well-built flow can handle most inbound call scenarios without a human, at a lower cost than a call center.
- SaaS Mode users packaging agent flows as a product. You can build an agent configuration into a snapshot and deliver it to clients as part of a white-label offer.
What to do next
If you have existing agent flows, open them in Agent Studio and check whether the new node types or branching options could simplify your setup. Use the test interface even on flows that already work, because it shows edge cases you may have missed.
For a full walkthrough, see the Agent Studio multi-step workflow build playbook. The Voice AI overview covers how Agent Studio connects to the voice layer.