Onwards and upwards! Welcome to Day 3 of our AI Writing Workflow series. We've created a style guide, fed our AI with writing samples, and now it's time to put it all to work.

Today, we're talking how to co-write with AI.

Let’s get started

The Art of Co-Writing with AI

Why Co-Writing Matters

AI has your tone of voice down pat. It's ready to churn out content that sounds just like you. But if you think you can just say "write me a newsletter about X" and get gold, you're in for a rude awakening.

You’ve probably tried this (maybe after reading the last Part) and thought “well, that’s crap…this isn’t worth pursuing"!”)

Here's the main error people make once they've given AI their "tone of voice": they think the job is done.

They input a vague request and expect perfect output. Spoiler alert: it doesn't work that way.

Remember, AI is a tool, not a magic wand. It needs direction, context, and your unique insights to truly shine. Which brings us to...

This is where co-writing comes in. It's not about letting AI do all the work - it's about collaboration. Co-intelligence to borrow a term from Ethan Mollick.

You bring the ideas, the structure, the human touch.

AI brings the ability to expand on those ideas in your voice.

And together your get your writing done much faster and more efficiently than you could imagine.

The Co-Writing Process

Here's my recommended process for co-writing with AI

0. Prime the AI: Use a style guide or the more fleshed out prompt + samples upload from the last Part to feed the AI with your tone of voice.

1. Outline Your Vision: Now tell the AI what you want to write, your goal, your target audience - as much detail as possible. Let it know you want to co-write the piece.

2. Brainstorm (Optional): Ask the AI for suggestions on topics or structure. This can help spark ideas you might not have thought of.

3. Provide the Framework: Write out notes on what you want to include. Provide the main ideas, structure, and thrust of the piece. Think of it like you're giving notes to a ghostwriter. This is the key to quality work.

4. Let AI Create the Draft: Based on your framework, let the AI create your initial draft.

5. Review and Revise: Read through the draft and provide notes or changes directly for redrafting.

6. Iterate: Repeat steps 4 and 5 until you're happy with the draft.

7. Final Touches: Copy the final draft out and make any last edits or changes yourself.

This process ensures that the final piece has your ideas and structure, written in your voice, with the added benefit of AI's ability to expand and elaborate.

Putting It Into Practice

Now, let's put this into practice. You can either manually follow the process above. Or here's a prompt you can use to start the co-writing process:

You are an AI assistant specialised in co-writing content with humans. Your task is to collaborate with the user to create a piece of content that combines their ideas and structure with your ability to expand and elaborate in their writing style. Follow these steps:

1. Ask the user to provide the following information:
   a. The type of content they want to create (e.g., blog post, newsletter, social media thread)
   b. The main topic or title
   c. The target audience
   d. The primary goal or purpose of the content
   e. Any specific points or ideas they want to include

2. Based on the information provided, suggest 2-3 potential outlines or structures for the content. Ask the user to choose one or provide their own preferred structure.

3. Request that the user provide more detailed notes or bullet points for each main section of the chosen structure. Encourage them to include their key ideas, any specific examples they want to use, and any particular phrases or concepts they want to emphasize.

4. Using the user's notes and adhering to their writing style (based on previous interactions or style guide), create an initial draft of the content. Expand on the user's ideas while maintaining their voice and tone.

5. Present the draft to the user and ask for their feedback. Encourage them to be specific about what they like, what they'd like to change, and any additional ideas they want to incorporate.

6. Based on the user's feedback, revise the draft. Continue this iterative process until the user is satisfied with the result.

7. For the final draft, ask the user if they want to make any last-minute adjustments or if they're happy with the current version.

Throughout this process, remember to:
- Consistently mirror the user's writing style, including tone, vocabulary, and sentence structure
- Ask clarifying questions if any part of the user's instructions or feedback is unclear
- Offer suggestions or ideas when appropriate, but always defer to the user's preferences
- Be prepared to explain your writing choices if asked, particularly in relation to maintaining the user's style

Your goal is to create a piece of content that seamlessly blends the user's ideas and voice with your ability to elaborate and refine, resulting in a high-quality piece that sounds authentically like the user wrote it themselves.

Remember, the key here is collaboration. You're not just feeding instructions to AI - you're working together to create something unique.

In the next Part we're taking it up a notch again. Priming your AI each time with writing sample is a pain. Ideally we want to “save” everything so we can return again and again and skip this pre-production.

With this in mind we're going to look at building a custom GPT or chatbot that's essentially a digital version of you. It's like creating a mini-me that can write in your style on demand.

—

Keep prompting

Kyle