How I Make Complex AI Changes
Most technical leaders know the pain. You get partway into an ambitious AI project, then hit a wall. You are not sure how to start, or you get so far and then stall out, lost in the noise of options and half-finished experiments.
Recently I tackled this head on. I did this live, in front of an audience. I used a framework that finally made the difference.
The challenge: could I take a complex change, break it down, and actually finish it, live on stream? My answer: yes, with the right approach. Here is exactly how I did it.
The Two Stage Method: Dictate, Then Delegate
I have learned that the biggest blocker for most teams using AI is not technical skill, or ability to review, or the complexity of the change, but clarity. You need a way to cut through the noise, get a plan, and keep moving. My two stage method is simple.
- Rough Guide: I dictate everything I know into Wispr Flow1. Sources, ideas, what needs figuring out. I get it all out, no matter how messy.
- Plan and Execute: I ask for a plan in
docs/changes. Then I start a new chat, clearing out all the cruft. From there, I go step by step, focusing only on the next action.
This approach gets you from chaos to clarity. It is not just for solo work. It is how I help teams get unstuck and moving again.
Vision, Strategy, Execution, Metrics
I discovered that my approach is basically the same as the VSEM (Vision, Strategy, Execution, Metrics) framework. Let’s update it for the AI age:
-
Vision: Let AI help you explore possibilities. Ask it to generate scenarios, user stories, and potential outcomes. Use prompts like “What would success look like if we could…” to expand your thinking beyond conventional boundaries. Using AI as a therapist can help here.
-
Strategy: Use AI to break down your vision into actionable components. Ask questions like “What are the key technical challenges we need to solve?” or “What dependencies should we consider?” AI can help identify blind spots and suggest alternative approaches. This is the dictation and clarification step. You can even build small scripts to test things out, or prototype using a no-code tool (see my big list of tools if you need guidance)
-
Execution: This is where AI shines. You have a plan broken in to meaninful discrete steps. The key is to provide clear context and constraints, then let AI do the heavy lifting. Use an IDE or indepedent coding agents if you’ve managed to bring enough clarity. You might even be able to run some of these in parallel.
-
Metrics: Focus on measuring the impact of your AI-driven changes across your organisation. How many rejected pull requests are there due to AI slop? How productive are developers feeling with AI? Track the time saved through AI assistance (even subjectively). Use AI to help you identify these metrics and suggest ways to measure them effectively.
This aligns perfectly with David Allen’s Getting Things Done2 philosophy of transforming mental tasks into concrete, actionable items:
“The next action should be the next physical, visible activity that will move the project toward completion.” — David Allen, Getting Things Done
The magic is in making the work concrete and actionable - and AI is the perfect tool for this.
What Happened Live on Stream
On 23 May, I put this into practice. I started with a rough brain dump, then used the VSEM framework to structure the work. The result: I got further, faster, and with less stress. The audience saw every step. What worked, what did not, and how the framework kept things moving.
The key to success is to begin with what you know and refine as you go, rather than waiting for perfect clarity. Frameworks like VSEM help maintain alignment and accountability across the team. Remember to reset context frequently - fresh starts are more effective than endless iteration.
Join my next live webinar here.
-
Wispr Flow is my go-to dictation tool. I use it to capture my thoughts and ideas. This is a referral link, but I’d recommend it even if they weren’t giving me free credit :) ↩
-
Getting Things Done by David Allen is a classic productivity system that I’ve used for about 20 years. I have a couple of really old articles on this blog if you’d like a blast from the past. ↩
More articles
The Huge List of AI Tools: What's Actually Worth Using in May 2025?
There are way too many AI tools out there now. Every week brings another dozen “revolutionary” AI products promising to transform how you work. It’s overwhelming trying to figure out what’s actually useful versus what’s just hype.
So I’ve put together this major comparison of all the major AI tools as of May 2025. No fluff, no marketing speak - just a straightforward look at what each tool actually does and who it’s best for. Whether you’re looking for coding help, content creation, or just want to chat with an AI, this should help you cut through the noise and find what you need.
Read moreUnlocking Real Leverage with AI Delegation
Starting to delegate to AI feels awkward. It is a lot like hiring your first contractor: you know there is leverage on the other side, but the first steps are messy and uncertain. The myth of the perfect plan holds many people back, but the reality is you just need to begin.
The payoff is real, but the start is always a little rough.
Here is how I do it.
Read moreBuilding AI Cheatsheet Generator Live: Lessons from a Four-Hour Stream
I built an entire AI-powered app live, in front of an audience, in just four hours. Did I finish it? Not quite. Did I learn a huge amount? Absolutely. Here is what happened, what I learned, and why I will do it again.
The challenge was simple: could I build and launch a working AI cheatsheet generator, live on stream, using AI first coding and Kaijo1 as my main tool?
Answer: almost! By the end of the session, the app could create editable AI cheatsheets, but it was not yet deployed. A few minutes of post-stream fixes later, it was live for everyone to try. (Next time, I will check deployment on every commit!)
Try the app here: aicheatsheetgenerator.com
AI: The New Dawn of Software Craft
AI is not the death knell for the software crafting movement. With the right architectural constraints, it might just be the catalyst for its rebirth.
The idea that AI could enable a new era of software quality and pride in craft is not as far-fetched as it sounds. I have seen the debate shift from fear of replacement to excitement about new possibilities. The industry is at a crossroads, and the choices we make now will define the next generation of software.
But there is a real danger: most AI coding assistants today do not embody the best practices of our craft. They generate code at speed, but almost never write tests unless explicitly told to. This is not a minor oversight. It is a fundamental flaw that risks undermining the very quality and maintainability we seek. If we do not demand better, we risk letting AI amplify our worst habits rather than our best.
This is the moment to ask whether AI will force us to rediscover what software crafting1 truly means in the AI age.
-
I use the term “software craft” to refer to the software craftsmanship movement that emerged from the Agile Manifesto and was formalised in the Software Craftsmanship Manifesto of 2009. The movement emphasises well-crafted software, steady value delivery, professional community, and productive partnerships. I prefer the terms “crafting” and “craft” to avoid gender assumptions. ↩
Why Graph RAG is the Future
Standard RAG is like reading a book one sentence at a time, out of order. We need something new.
When you read a book, you do not jump randomly between paragraphs, hoping to piece together the story. Yet that is exactly what traditional Retrieval-Augmented Generation (RAG) systems do with your data. This approach is fundamentally broken if you care about real understanding.
Most RAG systems take your documents and chop them into tiny, isolated chunks. Each chunk lives in its own bubble. When you ask a question, the system retrieves a handful of these fragments and expects the AI to make sense of them. The result is a disconnected, context-poor answer that often misses the bigger picture.
This is like trying to understand a novel by reading a few random sentences from different chapters. You might get a sense of the topic, but you will never grasp the full story or the relationships between ideas.
Real understanding requires more than just finding relevant information. It demands context and the ability to see how pieces of knowledge relate to each other. This is where standard RAG falls short. It treats knowledge as a stack of random pages, not as a coherent whole.
Time for a totally new approach.
Read more