Adjust a workflow with an AI assistant
A workflow is stored as a JSON file, which is plain text. That makes it something an AI assistant such as ChatGPT or Claude can read, explain and rewrite for you. You do not need to understand every node to make an example your own.
Before you begin
- Never paste credentials or API keys into an AI assistant. The downloads in the library are sanitized, but check anything you export yourself.
- Always test an adjusted workflow against a staging environment before production.
Adapt an example to your process
- Download the workflow JSON from its library page or copy your workflow from n8n.
- Paste it into the assistant with a prompt that describes your change. For example:
Here is an n8n workflow in JSON. Change it so it checks quotes
instead of orders and runs every Monday at 07:00 instead of daily.
Keep everything else the same and return the full updated JSON.
- Import the returned JSON into n8n.
- Reconnect your credentials. They are never part of the JSON.
- Run the workflow once manually and check the result before you activate it.
Convert an example to another tool
You do not have to use n8n. The JSON describes what the workflow does, so an assistant can rebuild it elsewhere:
Here is an n8n workflow in JSON. I use Make instead of n8n.
Explain what this workflow does step by step, then describe how
to rebuild it in Make, including the modules to use and the
settings per module.
The same works for Zapier, Power Automate or plain code. After converting, verify every step against the original guide on the entry's library page, since the assistant rebuilds the logic and details can shift.
Let the assistant explain before you change
If you only want to understand an example, ask for an explanation first:
Explain in plain language what this n8n workflow does, what data
it reads from the Propeller API and what it changes or creates.
This is a fast way to judge whether an entry fits your process before you invest time in it.