One of the biggest mistakes people make with generative AI is asking the AI to produce too much content at once. They ignore basic writing process steps thinking the AI can just “do it all for them.” There is an assumption that you can just give the AI a broad prompt and receive a lengthy, high-quality article in return.

In reality, this approach rarely works and often leads to disappointing results that need extensive editing or rewriting.

The issue is that most AI systems today have narrow capabilities (such as context length). They can do specific, focused tasks very well, but struggle with open-ended broad requests. When asked to write an entire piece from scratch, the AI does not have a strong sense of cohesion, flow, or direction. The output tends to be meandering, repetitive, or off-topic. As the human co-writer, you then have to do the heavy lifting to turn AI-generated ideas and draft text into something usable.

Instead, a much better approach is to use an iterative writing process. This is a method where the AI assists you in stages rather than bulk outputs. Have the AI focus on one section or part of the draft at a time.

With a narrowed scope, AI can produce higher-quality writing for the given prompt. Its outputs build upon each other, so the AI learns and improves with each iteration. You maintain more control over the direction of the content and can see how the draft is coming together – making changes as needed.

The Iterative Approach is More Effective

If you remember your writing process from high-school, the iterative approach will make a lot of sense to you.

An iterative approach to generative AI writing is the most effective method for using AI such as ChatGPT to write content. It allows the AI to:

• focus on one aspect at a time
• build upon its work with each iteration
• learn from feedback
• give you more control over the process,
• and results in a high-quality draft that requires minimal editing

While it may seem like more “work” up front, you’re actually producing significantly better content with less post-editing work.

It Focuses the AI’s Effort

When you focus the AI on one part of the draft at a time, it can do a better job with each specific task. The AI doesn’t have to split its capabilities between coming up with an outline, maintaining cohesion, and other aspects needed for a full long-form piece.

It can focus solely on the prompt for that iteration, whether that is developing an argument, writing an introduction, or adding supporting examples. Greater focus equals higher quality.

Each Iteration Builds Upon the Previous

With each iteration, the AI’s outputs also build upon each other. The AI maintains context from what it has already written to continue improving its writing over subsequent iterations.

As the human co-writer, you also provide feedback and input for the AI after each iteration. This helps further guide the AI and refine its writing. Over multiple iterations, the AI learns your preferences for style, tone, and content – and applies those lessons to the full draft.

You Have More Control

An iterative approach also gives you more control over the result compared to asking for a full draft at once. After each iteration, you can see how the draft is progressing and make any necessary changes to direction or focus before continuing.

Having more control means you remain the most valuable part of the process and the result is a reflection of you, not the AI.

Less Post-Editing Required

If the AI goes off on a tangent or includes irrelevant information, you can catch it right away. This is much better than getting an entire poorly-written draft and having to edit the entire thing. This back-and-forth process, where the AI writes and you review/refine, leads to a draft that matches your vision.

Using an iterative approach, you can leverage the power of generative AI for writing while maintaining the benefits of crafting a draft yourself. The AI becomes a collaborative co-writer rather than just a tool to produce a bulk of text that needs heavy revision.

The result is greater efficiency and higher quality work that reflects your original goals and vision.

Writing Process Steps for ChatGPT Example

Let’s go over an example of how the iterative approach would work if we wanted to create an article or blog post. I’m going to give you some basic structure of how I involve the AI in my writing process.

Step 1. Come up with the topic

Let’s say you want to write an article about the rise of AI and its impact on jobs. Begin by having the AI brainstorm talking points for you. Your first prompt would be something like the following:

Prompt:
"I need to write an article about the rise of AI and its impact on jobs. What are some important points I should cover in an article on this topic?"

Wait for the AI’s response and read what it comes up with. The AI may respond with points like:

• Advancements in machine learning enabling new AI capabilities
• AI automating and replacing some routine jobs
• Growth of new jobs in tech and AI fields
• The need to educate and retrain workers for new types of work.

You can ask the AI to add, subtract, or modify points if you want. Once you have all the talking points, then move to the next step. AI is pretty good at spelling and grammar, so you shouldn’t have to worry about that much. Just ensure it’s writing every idea that you want included.

Step 2. Ask the AI to produce an outline

Now we want to have the AI produce a workable outline based on the topic + talking points. Since it already knows what all of that is, our prompt can be very simple:

Prompt:
"Please create an article outline based on our topic and talking points we've decided on."

The AI might generate an outline like the following:

• I. Introduction to the growth of artificial intelligence
• II. AI and automation of routine jobs like customer service reps, cashiers, and factory work
• III. Rise of new technology jobs in AI, engineering, and computer science
• IV. Need for education and job retraining programs to prepare workers
• V. Examples of companies investing in AI and the future of work
• VI. Conclusion: The future of human-AI collaboration in the workforce

Again, work with the AI as needed to modify the outline to your liking. Make sure all the main points are the ones you want included in the final published article. Once it’s where you are happy with it, move to the next step.

Step 3. Have the AI draft the first section

Now it’s as simple as telling the AI to produce the first section. Your prompt can be as simple as the following:

Prompt:
"Please write section I."

Since it knows the context, it will understand exactly what section you’re referring to, and what that section should contain. Review its output and edit as needed, just as you have with other steps.

You can ask ChatGPT to rewrite a single sentence or an entire paragraph. Prompt it just like you would a real human writer.

Step 4. Repeat the process with the remaining sections

Have the AI draft each section, one at a time, reviewing and revising each draft section as needed. Provide feedback to improve the AI’s writing over time.

Step 5. Review everything

Once all sections are complete, review the full draft article for flow and coherence. Make any final edits. This work should be extremely minimal since you’ve effectively done the proofreading along the way.

These writing process steps allow you to break a large task like writing an article into smaller parts. The AI then assists you in drafting each section. You control the overall process and output. The key is providing feedback at each step to guide the AI in producing work that matches your goals and quality standards. With multiple iterations, the AI will learn and improve.

The Best ChatGPT App for the Iterative Approach

Using the iterative approach can work just fine in an app like ChatGPT. But it’s not the best way since you’ll likely be building the long-form content in some type of document.

That is why, when I built Magai, I built a Document editor that you can have open side-by-side with the Chat. This way you can have your conversation with the AI and build your entire article without having to switch between tabs or apps.

You can execute each of the above writing process steps without ever switching to a different tab or app.

When you’re done, you can copy the entire document text with a single click and paste it wherever you need it to be.

Magai also gives you loads of features to help you organize your chats, collaborate with team members, and more.

If you haven’t yet, give Magai a try today! You get 7 days and 10,000 premium words free to test-drive the app.

Conclusion

Using these proper writing process steps with generative AI leads to greater efficiency, higher quality results, and a better overall writing experience.

It may seem more convenient to get a full AI draft at once. However, an iterative approach will save you time and lead to a final product you can be happy with in the end. The best stories are crafted, not commanded into existence.