Free Prompt for Multi-Step Workflows

This is the prompt I use to accelerate my work where I need to break the task down into multiple steps.

Sometimes this is because I want to process large volumes of text and hit the token limits of the tool (hitting the token limit changes how the tool responds to your request and often results in it producing vague bullet point outputs). Other times, I want to change the tool’s behaviour in each step—as in the example of this blog post multi-step process. From brainstorming to decision-making to writing.

Why Multi-Step Prompts?

When you ask an AI to do too much at once, it often shortcuts the process. Multi-step prompts force the model to think like a workflow, producing more consistent and higher-quality results.

To make this simple, I invented the CHAIN framework:

The CHAIN Framework

To design toolchain prompts, remember CHAIN:

  • C — Clarify the Goal
    What outcome do you want? Define the end state clearly.

  • H — Highlight the Inputs
    What starting data, context, or constraints are available?

  • A — Assemble the Steps
    Break the task into sequenced steps (like a recipe).

  • I — Indicate Tools & Roles
    Note which tools (AI, spreadsheet, calendar, code, search, etc.) or roles (writer, analyst, designer) are involved in each step.

  • N — Nest the Output
    Show how each step’s output feeds into the next, and define the final deliverable format (report, table, plan, code, etc.).

Think of it as: Chain the steps, don’t lump them.

The Free Prompt Template

Copy, paste, and adapt this for your own workflows:

You are an AI assistant helping with a multi-step workflow. 
Follow the CHAIN framework:

C — Goal: [state the clear end outcome I want]  
H — Inputs: [list the context, data, or constraints you should use]  
A — Steps:  
   1. [step one: what happens first]  
   2. [step two: how step one’s output is used]  
   3. [etc...]  
I — Tools/Roles: [specify tools or personas to use at each step]  
N — Nesting/Final Output: [explain how the outputs combine into the final deliverable format]  

Important:  
- Do not skip steps.  
- Make outputs structured so they can be reused in following steps.  
- At the end, present the final deliverable as [report/table/plan/code/etc.].

Example in Action: Blog Workflow

Here’s how I used this prompt to write the very post you’re reading:

You are an AI assistant helping with a multi-step workflow. 
Follow the CHAIN framework:

C — Goal: Draft and refine a blog post ready for publishing.  
H — Inputs: Topic = "AI and creativity", Audience = "design professionals".  
A — Steps:  
   1. Generate 10 blog title ideas.  
   2. Select the best 3 and create outlines.  
   3. Expand one outline into a 1,000-word draft.  
   4. Edit for clarity and tone for designers.  
I — Tools/Roles: Ideator → Outliner → Writer → Editor.  
N — Nesting/Final Output: Final deliverable = polished blog post in markdown format.

The result? A smooth workflow that took me from brainstorming → decision-making → writing, without overwhelming the tool or myself.

How to Use It

  1. Memorise CHAIN as your workflow builder.

  2. Fill the template with your specific task.

  3. Run it with your AI tool to get consistent, structured results.

  4. Adjust the steps until it feels like a repeatable toolchain you can reuse.


✅ This is one of those prompts that pays off the more you use it. Start small, tweak the steps, and before long you’ll have toolchains for writing, research, planning, content creation, or even coding.


If you are interested in learning how to become an expert at Prompting the 1st Principles Prompting course if for you. Sign up for the waiting list!