Most people use AI the way they use a search engine — one question at a time, starting from scratch every session. That's the most expensive way to be unproductive with AI. The real unlock isn't a better prompt. It's a better process.
This lesson teaches you how to codify your workflows — to stop reacting and start building reproducible, AI-augmented systems that deliver consistent, high-quality output every time.
If you do something repeatedly, spend significant time on it, and add little unique human value to the execution — that's a workflow waiting to be codified. The goal isn't to replace your judgment. It's to stop wasting your judgment on tasks that don't deserve it.
AI LESSON
Codify Your Workflows: The AI Lesson Most People Skip
Stop Reacting. Start Systematizing.
I publish four newsletters reaching over 250,000 subscribers. For months, I was doing it the hard way — going back and forth with AI tools, re-explaining my preferences every session, manually formatting, proofreading with separate tools like Grammarly, and spending hours on work that didn't require my expertise.
The problem wasn't the AI. The problem was me. I wasn't giving enough information up front to get to a final draft. I wasn't codifying what I already knew about my own process.
Here's what I was doing wrong, and what I changed:
Before (Ad Hoc):
Re-explained style preferences every session
Used Grammarly + manual proofreading as separate steps
Featured images required separate tools and back-and-forth
Admin couldn't help effectively — my knowledge was in my head
8–10 hours per week on newsletter production
After (Codified):
Complete style guide loaded into AI Projects — em dashes, Oxford comma, banned words, voice patterns
LLM handles style, grammar, and formatting in one pass using the style guide
Manus calls NanoBanana to generate images, charts, and infographics from my content
SOPs document exactly where AI acts and where humans review
~2 hours per week with higher consistency
The Lesson: A Four-Step Framework
The lesson walks you through a four-step process for codifying any repetitive workflow. This maps directly to the S.M.A.R.T. Framework (Sort, Match, Automate, Refine, Take Control) that you may already know from my previous lessons.
Step One: Audit Your Repetitive Work
Look at everything you do repeatedly that takes significant time and where you add little unique human value. For me, that meant formatting, proofreading, and creating images for newsletters. This is the Sort step in the S.M.A.R.T. framework, where you categorize tasks as Repetitive, Data-Driven, or Creative/Strategic.
Structured Interview Technique:
Instead of staring at a blank page trying to list your tasks, have AI conduct a structured interview. Tell Claude or ChatGPT: Ask me questions about my weekly workflow, one at a time, to help me identify tasks I should codify. This lets you spend your valuable thinking on the answers, not on figuring out the right questions.
Step Two: Build Your Knowledge Base Up Front
The biggest mistake people make with AI is not giving enough information up front. Every time you re-explain a preference, a style choice, or a process step, you're burning time that should have been invested once.
For my newsletters, this meant creating:
A comprehensive style guide — em dashes with spaces, Oxford comma, banned words (demystify, tapestry, landscape, delve), voice patterns, number formatting, and headline conventions
Reference examples of published newsletters that match the target voice
A template structure with section patterns and closing signatures
Image generation guidelines with brand-consistent parameters
Step Three: Choose Your Container and Codify
Take the repetitive task and put it into a persistent AI container. The three main options:
ChatGPT Projects: Best for team workflows, persistent context across sessions, and file management. Key advantage: Custom GPTs + Projects combo for behavioral control and API integration.
Claude Projects / Skills: Best for deep analysis, long-form content, and style-sensitive writing. Key advantage: Reusable skills that persist; strong at matching voice and nuance.
Manus: Best for end-to-end production, including image generation and multi-tool orchestration. Key advantage: Can call external tools like NanoBanana for images, charts, and infographics.
Step Four: Create the SOP — Human + AI Handoffs
The final and most important step is documenting exactly where AI should act and where humans must intervene. This is your Standard Operating Procedure — and it's what makes the workflow transferable. Without it, the process lives in your head, and you're back to being the bottleneck.
Example: My Newsletter Production SOP
Ideation (Human — Mark): I come up with the topic, the angle, and the key points I want to make. This is where my thirty years of technology experience and editorial judgment matter most.
Information Assembly (Human — Mark): I gather the sources, data, examples, and personal anecdotes I want to include. I decide what the reader needs to know.
Structured Interview (AI): Optionally, I have Claude interview me about the topic to extract thinking I might not have articulated. This saves me from staring at a blank page.
Organization and Gap Analysis (AI): Claude organizes the raw material, identifies holes in my argument, suggests missing evidence, and proposes a structure.
Review and Augment (Human — Mark): I review the organized draft, fill in gaps the AI identified, and provide any additional context or direction.
Polish and Format (AI): The AI writes the final draft using my complete style guide — handling formatting, grammar, tone consistency, headline conventions, and image generation in one pass.
Editorial Oversight (Human — Admin): My admin provides a fresh set of eyes — checking for accuracy, readability, and anything that doesn't sound like me.
Final Quality Check (Human — Mark): I do a final review, approve, and publish.
The Result: Newsletter production went from 8–10 hours per week to approximately 2 hours — with higher consistency, fewer errors, and a process my admin can execute independently when I'm traveling or speaking.
Your Action Items for This Week
Pick one workflow you do every week that takes more than an hour and where you're mostly executing, not thinking.
Document your preferences for that workflow. Write down every decision you make on autopilot — formatting choices, tone preferences, structural patterns, recurring instructions.
Load it into an AI Project. Choose ChatGPT Projects, Claude Projects, or Manus. Give the AI your complete context up front.
Run it once and note the gaps. Where did the AI miss? What did you have to re-explain? Add that to your knowledge base.
Write the SOP. Document which steps are human and which are AI. Make it transferable — so someone else can follow it.
Bonus: Structured Interview Prompt
Include this as a copy-paste prompt readers can use immediately:
I want to codify one of my repetitive workflows using AI. Interview me to help identify the best candidate. Ask me one question at a time about: (1) What tasks I do every week, (2) How long each takes, (3) Where I add unique judgment vs. just executing, (4) What preferences or style choices I make on autopilot, and (5) Whether someone else could do this task if they had my instructions. After the interview, recommend which workflow to codify first and outline the knowledge base I'd need to create.
AI Extra Credit
Keep learning with these upcoming free virtual events from the All Things AI community.
February 24 | The Missing Link: Adding Your Data to Your App — Learn how to connect vibe-coded front-end apps to real data using Make.com webhooks, data mapping, and multi-step automation workflows.
March 3 | How Small Orgs and Non-Profits are Getting Value out of AI — Real-world AI adoption strategies for small organizations and nonprofits, including ready-to-adapt prompts and process maps you can bring back to your team.
March 10 | The Path to Becoming AI-First: An Operator’s Playbook — This session explores what it really means to become AI-first: starting with process, increasing leverage, and learning how to manage digital employees alongside human employees.
IN PARTNERSHIP WITH ALL THINGS AI
All Things AI 2026 — March 23–24 | Durham Convention Center, NC
I produce the All Things AI Conference with my business partner, Todd Lewis, founder of All Things Open. We are committed to upskilling and aim to deliver the most valuable, accessible expert-led workshops in the industry. Here’s what’s on tap in Durham in March. Workshops sold out in 2025. Don't wait. Check out all the workshops here.
Conference Pass — $199 — Tuesday, March 24. Full conference access, 50+ sessions across 4 tracks, networking events, and session recordings.
AI for DevOps Workshop + Conference — $299 — Monday–Tuesday, March 23–24. Full-day hands-on workshop with John Willis (Author of the DevOps Handbook and co-founder of the DevOps movement) plus full conference access.
AI for Business Workshop + Conference — $299 — Monday–Tuesday, March 23–24. Full-day hands-on workshop with Mark Hinkle plus full conference access.
AI for Agents Workshop + Conference — $299 — Monday–Tuesday, March 23–24. Full-day hands-on workshop with Don Shin plus full conference access.
Prices increase after March 17. Compare that to $1,000–$3,000+ at other AI conferences.
I appreciate your support.
Your AI Sherpa,
Mark R. Hinkle
Publisher, The AIE Network
Connect with me on LinkedIn
Follow Me on Twitter
