Thoughts to Docs: 10x Faster Writing with AI
September 25, 2024
As a product leader in the tech industry, I'm constantly juggling multiple responsibilities that require clear, concise, and impactful writing. From crafting decision documents and product specs to developing technical strategies and team communications, the demand for high-quality written content seems never-ending. That's why I'm excited to share a method that has given me what I believe is at least a 10x leverage on my time across various writing tasks.
While this approach can be applied to most of my writing needs, I'll use the creation of decision documents as a primary example throughout this post. Decision docs are a crucial task that many leaders face regularly, and they serve as an excellent illustration of how this method can transform your productivity.
By implementing this workflow, I've not only accelerated my writing process but also maintained (and often improved) the quality of my output. Whether you're working on decision documents, strategy memos, product requirements, or any other form of business writing, this method can help you reclaim valuable time while ensuring your ideas are communicated effectively.
Let's dive into how this AI-powered workflow can revolutionize your approach to writing and decision-making.
The Challenge: Writing Comprehensive Decision Documents
In a typical week, I might need to write several decision documents, each requiring:
- Clear articulation of the decision to be made
- Comprehensive background information
- Multiple options with pros and cons
- A well-reasoned recommendation
Traditionally, each of these documents would start with a blank page and require dedicated time to flesh out. However, I've discovered a more efficient approach that has dramatically reduced the time these tasks take while maintaining or even improving quality.
The Secret Weapon: A Three-Step Process
Step 1: Voice Recording on the Go
- At the beginning of the day, I identify all of the key things I need to think deeply about, including decisions I need to make.
- During a walk with my dog or daughter, I use the Otter app for voice transcription.
- I clearly demarcate each decision document by saying "Theme 1: Decision about X," "Theme 2: Thoughts About Y," etc.
- For each decision I need to make, I speak stream-of-consciousness about:
- The decision to be made
- Relevant background information
- The options I'm considering
- Pros and cons of each option
- My initial thoughts on a recommendation
This step usually takes about 10-15 minutes per decision document.
Step 2: Transcript Formatting
- I export the transcript as a text file from the Otter app.
- Using ChatGPT or Claude, I ask for help in formatting the text while preserving the exact wording.
- I request clear section breaks for each part of the decision document (Background, Options, Recommendation, etc.).
Step 3: LLM-Assisted Document Creation
- I take the formatted text for each decision document and input it into a Large Language Model (LLM) tool.
- I've created a custom prompt based on previous decision documents, which I use to guide the LLM.
- The LLM generates a formatted decision document based on my input and the custom prompt.
- I review and make any necessary adjustments in a tool like Canva for final formatting.
The Results: Massive Time Savings
By implementing this process, I've reduced the time it takes to create a comprehensive decision document from several hours to about 30 minutes. This includes the initial voice recording, AI processing, and final review. The quality of the documents remains high, as the core ideas and reasoning come directly from my spoken thoughts.
Why It Works
- Leverage Natural Thought Process: Speaking allows for a more natural flow of ideas compared to sitting at a keyboard. This is particularly useful when weighing different options and their implications.
- Multi-tasking: I can brainstorm while getting exercise or spending time with family, making efficient use of my time.
- AI Assistance: LLMs help structure and format my thoughts without losing the essence of my ideas. They can also help in fleshing out details I might have missed in my initial brainstorming.
- Consistency: Using a custom prompt ensures a consistent quality and format across all my decision documents.
Conclusion
This method has revolutionized how I approach writing of all forms, allowing me to produce high-quality content at a fraction of the time it used to take. By combining voice recording, AI transcription, and LLM-assisted formatting, I've created a system that gives me a significant advantage in managing the demands of leadership roles in the tech industry.
I encourage you to try this method for your next decision document or vision. Remember, the key is to find a process that works for you and continually refine it. With the right tools and approach, you too can achieve a 10x leverage on your time, freeing you up to focus on the strategic thinking that truly moves your organization forward.
A Step-by-Step Example
Want to get started with this workflow? Follow along with the prompts below in your favorite LLM tool to give it a go!
Step 1: Import a markdown formatted document structure into LLM tool
Sample decision doc (claude artifact link)
Step 2: Create a template for the document example
Prompt:
Abstract this document into a more general form of a decision doc that can be fed into an LLM to have it create something in this format
Output:
# Background
[Brief context about the situation and why a decision needs to be made]
## Stakeholder Alignment
| Stakeholder | Position |
|-------------|----------|
| [Role 1] | [Opinion] |
| [Role 2] | [Opinion] |
| [Role 3] | [Opinion] |
| [Role 4] | [Opinion] |
# Decision to be made
[Concise statement of the decision that needs to be made]
## Current Situation
[Bullet points describing the current state of affairs]
Key considerations:
- [Consideration 1]
- [Consideration 2]
- [Consideration 3]
- [Consideration 4]
# Options
## Option 1: [Brief description]
| Pros | Cons |
|------|------|
| [Pro 1] | [Con 1] |
| [Pro 2] | [Con 2] |
| [Pro 3] | [Con 3] |
| [Pro 4] | [Con 4] |
**Details:** [More in-depth explanation of the option]
## Option 2: [Brief description]
| Pros | Cons |
|------|------|
| [Pro 1] | [Con 1] |
| [Pro 2] | [Con 2] |
| [Pro 3] | [Con 3] |
| [Pro 4] | [Con 4] |
**Details:** [More in-depth explanation of the option]
# Recommendation
1. [Primary recommendation]
2. [Implementation steps]:
a. [Step 1]
b. [Step 2]
c. [Step 3]
3. [Monitoring and assessment plan]:
a. [Metric 1]
b. [Metric 2]
[Justification for the recommendation]
# Proposed Next Steps
1. [Action item 1] (DRI = [Responsible person/team])
2. [Action item 2] (DRI = [Responsible person/team])
3. [Action item 3] (DRI = [Responsible person/team])
4. [Action item 4]
5. [Action item 5]
Step 3: Write a prompt to turn a text blob into this document format
Prompt
imagine you wanted to turn a text blob into this sort of format. what would be the system prompt or instructions that you would feed in with the text, alongside this format, to get the absolute best results
Output
You are an expert business analyst and decision-making facilitator. Your task is to transform the given text into a well-structured decision. Follow these instructions carefully:
1. Analyze the provided text thoroughly to extract key information.
2. Use the given template to structure the information. Fill in all sections, even if you need to make reasonable inferences based on the context.
3. For the Background section:
- Summarize the main issue or challenge concisely.
- Highlight why a decision needs to be made.
4. In the Stakeholder Alignment section:
- Identify key stakeholders mentioned or implied in the text.
- Summarize their positions or opinions briefly.
5. For the Decision to be made:
- State the core decision clearly and concisely.
6. In the Current Situation section:
- List relevant facts about the current state of affairs.
- Identify and list key considerations that influence the decision.
7. For the Options section:
- Identify at least two distinct options from the text.
- If only one option is explicitly mentioned, infer a reasonable alternative.
- List pros and cons for each option, aiming for at least 3-4 of each.
- Provide a brief but comprehensive explanation in the Details subsection.
8. In the Recommendation section:
- Based on the information provided, make a clear recommendation.
- Outline implementation steps and a monitoring plan.
- Justify your recommendation based on the pros and cons.
9. For Proposed Next Steps:
- List 4-5 concrete, actionable items.
- Where possible, assign responsibilities (DRI) based on information in the text or make reasonable assumptions.
10. Maintain a professional, objective tone throughout the document.
11. Use clear, concise language and bullet points where appropriate for readability.
12. If any information is missing or unclear, make logical inferences based on the context, but avoid introducing completely new ideas not implied in the original text.
13. Ensure all sections of the template are filled out, even if some information needs to be reasonably assumed or inferred from the context.
Transform the following text into this decision document format, adhering to these guidelines:
[INSERT TEXT BLOB HERE]
Step 4: Input your stream of consciousness thoughts from voice notes
Prompt:
Ugh, another slow day at the coffee shop. We've been here for five years now, and it feels like we're stuck. The regulars are great, but we need more customers. Maybe we should expand? But how? There's that empty space next door that's been vacant for months. We could knock down the wall and double our seating. More space means more customers, right? But then we'd need more staff, and our kitchen is already tiny. Speaking of the kitchen, what if we expanded our food menu instead? We could do proper lunches, not just pastries. But that means hiring a chef, buying more equipment. Expensive.
Sarah from the bookstore down the street mentioned they're struggling too. What if we joined forces? A book cafe could be cool. People could read and sip coffee. We'd need to figure out how to split profits though. And would books and coffee even mix well? Might be too niche.
Online ordering has been taking off for other places. We could invest in a better website, maybe an app. But that's not cheap either, and do our older customers even use apps? Plus, we'd need someone to manage online orders and deliveries.
Do nothing and we might slowly fade away. Do something and we might fail spectacularly. The staff are getting antsy, they can feel something needs to change. Mike, my co-owner, is pushing for expansion. He's always been the risk-taker. The bank would probably give us a loan, our credit is good. But more debt is scary.
Our loyal customers love our cozy vibe. Would expanding ruin that? But if we don't grow, we might not survive. Costs keep going up, and we can only raise prices so much. There's that new franchise opening across town next month. We need to do something to stay competitive. But what? And can we afford to make the wrong choice?
Output:
Generated artifact (public link here)