Who This Is For / When to Use
Use this feature when you want workflows to generate or analyze text dynamically, such as drafting messages, summarizing conversations, translating content, or classifying sentiment.
Available GPT Models
Workflow AI supports the following models:
GPT‑5 – Best for complex reasoning and advanced text generation
GPT‑5 Mini (default) – Balanced performance and cost for most workflows
GPT‑5 Nano – Fastest option for lightweight tasks
GPT‑4 Turbo – Strong performance for structured and long-form text
GPT‑4o – Optimized for advanced language understanding
GPT‑4o Mini – Cost‑efficient GPT‑4 option
The default model is GPT‑5 Mini.
Add the GPT Powered by OpenAI Action
Open the workflow editor.
Click Add Action.
Search for GPT powered by OpenAI.
Select an Action Type
Action Type defines what the AI will do.
Pre‑Built Action Types
Use these when you want standard behavior without writing a custom prompt:
Analyze Text Sentiment
Summarize Text
Translate Content
Custom Action Type
Use Custom when you want full control over the prompt and output.
Write your own instructions
Insert workflow variables
Control formatting and tone
Configure the Prompt
The prompt tells the AI exactly what to do.
Be explicit and concise
Include variables such as contact fields or conversation content
The AI only sees what is included in the prompt
Example prompts:
Lead follow‑up: Draft a friendly message to {{contact.name}} thanking them for their interest in {{contact.custom_value.service}}.
Conversation summary: Summarize this conversation in 2–3 bullet points: {{conversation.history}}
Email draft: Write a persuasive email promoting {{contact.custom_value.offer}}.
Enable Chat History (Optional)
When enabled, the model can reference previous executions of the same GPT action for the contact.
Chat History is stored per contact
Turning this on increases context size
Required if you want responses influenced by prior GPT outputs
Save the Action
After configuration, click Save Action to activate the GPT step in your workflow.
Common Issues and Fixes
AI Output Is Missing Context
Confirm required variables are included in the prompt
Conversation data is not included unless explicitly mapped
Output Is Too Generic
Switch to Custom action type
Add formatting rules or constraints to the prompt
Slow Execution
Use GPT‑5 Nano for lightweight tasks
Reduce prompt length when possible
Important Notes
This is a premium workflow action and may incur additional charges
The AI only processes the text provided in the prompt
Model choice directly affects cost, speed, and output quality