The AI Agent workflow action lets you automate multi-step tasks inside Captivation Hub Workflows using plain-language instructions and the tools you choose. Rather than building each step manually, you configure a single AI Agent action that decides what to do, which tools to call, and what inputs to pass at runtime. This guide walks through how the action works, how to set it up, and how to review its output.
Note: This is a premium action. Each execution will incur additional charges.
What is the AI Agent Action in Workflows?
AI Agent is a Workflow AI action that plans and executes tasks autonomously based on your instructions and the tools you make available. Instead of chaining several actions together yourself, the agent reasons through the goal, decides which tools to use, and runs them in sequence.
Key Benefits of the AI Agent Workflow Action
- Fewer hand-built steps — one AI Agent action can replace several rigid workflow branches.
- Adaptive decisions — the agent picks the right tool based on the situation rather than following a fixed path.
- Cross-channel context — built-in conversation history pulls in past messaging exchanges when relevant.
- Full traceability — execution logs show every model call, tool execution, and token usage detail.
AI Agent Action Details
These settings shape how the AI Agent behaves and provide the core inputs you'll need to configure the action.
- Action Name: Add a custom name. The default is "AI Agent" — renaming it to match the use case keeps your workflow easier to manage.
- Template: Choose a pre-built template that pre-fills the Instructions and Tools, or pick "Build Your Own" to start from scratch. Some templates require a specific workflow trigger.
- Instructions: Define what the agent should do, how it should make decisions, and any rules to follow. This is the most important part of the configuration.
- Enhance Prompt: Restructures loosely written instructions into a more detailed, organized prompt. You can undo the enhancement at any time.
- Model: Select the AI model that powers the agent.
- Tools: Choose the actions the agent is allowed to execute.
- Conversation Memory: Optional rolling summary of prior runs for the same contact.
Templates
Templates are pre-built agent configurations that pre-fill Instructions and Tools so you can get up and running quickly.
Writing Instructions
Instructions tell the agent what you want it to do. Clear, specific instructions consistently produce better results — describe the goal, the inputs, and any rules or edge cases.
Configuring Tools
Tools are the actions the agent is permitted to take during a run. For each tool, you can either let the AI decide the field values or set them manually.
Built-in Conversation History (Cross-Channel)
The AI Agent can pull in a contact's recent messaging history across channels when it's relevant to the task. There's nothing to configure — this is available by default.
What the agent can retrieve:
- Channel (where each exchange occurred)
- Message content
- Direction (inbound vs. outbound)
- Timestamps
- An AI-generated summary for high-level context
Channels covered: SMS, Email, WhatsApp, Instagram, Facebook Messenger, TikTok, and other connected messaging channels.
Step 1 — How to Set Up the AI Agent Workflow Action
A structured setup helps you choose the right template, write clear instructions, and equip the agent with the tools it needs.
- Go to Automations > Workflows.
- Create a new workflow or open an existing one.
- Add a relevant trigger such as Form Submitted or Contact Created, or pick a pre-built template during action setup.
- Click + Add or select the "Please select action" placeholder on the workflow canvas.
- Choose AI Agent from the Workflow AI category.
- Write your Instructions, then add the Tools the agent is allowed to use.
- Configure each tool — decide whether the AI should choose field values or whether you'll set them manually.
Or pick a Template — these populate the Instructions and Tools for you. If a template requires a specific trigger and that trigger isn't already in the workflow, the template will appear unavailable until it's added.
Advanced Options
Open Advanced Options if you'd like to enable Conversation Memory or configure Output Format. These give you more control over how the agent remembers prior runs and how its output is exposed to later workflow steps.
Conversation Memory: Off by default. When enabled, the agent uses a rolling summary so it can keep continuity across repeated runs for the same contact.
Output Format:
- Text — describe what the response should contain.
- JSON — return structured data based on a schema you define. You can add properties with a name, type, and optional description so later workflow steps can reference specific values.
Save the action when you're done.
Execution Logs
Execution Logs make it easy to see exactly what the agent did during a workflow run — useful for validating behavior, checking tool usage, and troubleshooting unexpected results.
After the workflow runs, open the Execution Logs tab in the workflow builder and pick an AI Agent execution to view the trace.
The trace can include:
- Agent Input — the full input received by the agent.
- LLM — each model call used to reason about the next step.
- Tool executions — each tool the agent called, along with execution time and success/failure status.
Each step can be expanded to review input and output data. The log view supports both a table format and a raw JSON view.
Pricing
The AI Agent is a premium action — usage incurs additional charges per execution based on the model selected and the tokens consumed. Check your account for the most current rates.
Frequently Asked Questions
Q: Can I use the AI Agent with any trigger?
Yes — the AI Agent works with any workflow trigger. Some templates suggest specific triggers because their instructions are written with that context in mind.
Q: What does Enhance Prompt do?
Enhance Prompt rewrites your instructions into a more organized, detailed prompt that includes clearer steps and edge-case handling. You can undo the enhancement at any time.
Q: What context does the AI Agent use when running in a workflow?
It uses your instructions, the tools you've enabled, and the contact and workflow data available at runtime. When messaging context is relevant, it can also pull the contact's conversation history across channels.
Q: What's the difference between Conversation Memory and Conversation History?
- Conversation Memory — a rolling summary of prior runs of this AI Agent step for the same contact (optional).
- Conversation History — cross-channel messaging history retrieval, used when needed (available by default).
Q: When should I use the JSON output format?
Use JSON when later workflow steps need to reference specific values from the agent's response — qualification results, structured data fields, that kind of thing.
Q: How can I see what the agent did during a workflow run?
Open the Execution Logs tab in the workflow builder and select the AI Agent execution. The trace includes model calls, tool executions, inputs, outputs, and token usage.
Q: Do templates lock me into a specific configuration?
Not at all. Templates pre-fill Instructions and Tools, but you can edit the instructions, swap models, add or remove tools, and adjust advanced options afterward.
Q: When should I enable Conversation Memory?
Enable it when the same contact may go through the agent step more than once and the agent should adapt based on prior runs. For one-time workflows, it's usually unnecessary.
Comments
0 comments
Please sign in to leave a comment.