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.
Why AI Delegation Feels Awkward
The first time I tried delegating to an AI agent, it felt exactly like onboarding a new team member. I was not sure what to hand over or how to check the work. The results were messy and there was a lot of copy/pasting. It took a lot longer than doing it myself.
Delegation is always awkward at the start, whether you are working with people or machines. The learning curve is part of the process, and the discomfort is a sign you are moving forward.
Here are five principles for getting started:
Start Small
Begin with a simple, concrete task. Even if it feels trivial, this gives you a clear baseline. The first attempt will be clumsy - that is normal. Each iteration helps you refine your instructions and identify potential issues.
Expect to Iterate
The journey to effective AI delegation is not linear. Each round improves your ability to specify requirements and spot problems. Do not expect perfection on the first try. Instead, focus on learning from each attempt.
Use Human Delegation Frameworks
Apply the same frameworks you use for human delegation. The VSEM framework helps separate vision, strategy, and execution. A GTD approach breaks work into manageable chunks. I wrote more about how to apply these delegation frameworks in AI coding in this post.
Treat it Like Hiring Contractors
Approach AI delegation like hiring contractors: experiment quickly and move on if something is not working. Do not get stuck trying to make a suboptimal approach work. The goal is to find what works for you.
Embrace the Mess
The path to effective AI delegation is not smooth. It is fiddly and sometimes frustrating. But every experiment brings you closer to real leverage. The mess is part of the process - embrace it and keep moving forward.
How to Delegate a Workflow
Start with a Fixed Task
Begin with something entirely fixed, without AI, to get used to the tools you have available.
For example, I use n8n to automate a Kit tag addition that creates a Google Calendar entry for my webinars. It was fiddly to create, but it helped me learn how the workflow tool worked.
If you are more technical, try using n8n. If you are not, Zapier is a great place to start. There are plenty more tools out there if you need more options.
Look for Decision Points
Once you are comfortable with the tools, look for decision points in your workflow where you would not normally give the job to a computer. Try and find a really simple decision that you make all the time.
If you are not sure what an AI might be able to do, ask an AI agent to brainstorm where an agent could help within your workflow. This kind of meta-thinking about strategy is a really important skill to develop in the age of AI.
Automate One Thing
Pick a small decision to automate. Set up the automation in your chosen tool. Add an AI agent to handle the decision point. Test, iterate, and refine.
Each experiment will teach you something new about the tool and what AI can do, slowly increasing your leverage over time.
If something doesn’t work, revisit when a new model comes out. Those who will be at the forefront of this will be the ones who are willing to go back and push the boundaries again, and experiment and learn. The tech is moving so fast that what didn’t work three months ago might just work now, unlocking big productivity gains.
An Example
Here is an example. Imagine you receive inbound emails asking for advice. You probably already automate parts of this process, using rules, by marking particular emails as important based on who they are from. Try first to move this process to a workflow tool. (Note: this could get expensive if you get a lot of email, so perhaps do the first pass of filtering via your existing email rules, such as emails not directly addressed to you.)
Once that’s done, how do you process the rest? Normally, you might read each one and decide: should you reply personally, refer the person to a resource, or have AI draft a first response?
This is a perfect decision point to automate. Set up a workflow where AI reviews the email and drafts a reply for you to approve, or marks the email as important for your attention. You can start by automating the triage step, then gradually hand over more of the process as you gain confidence.
The Payoff and Next Steps
Delegating is always awkward. It could be the job is done only 80% as well as if you did it. But if we learn to delegate and let go of our perfectionism, real leverage opens us for us. The beauty of the AI age is that this leverage is beginning to be available to all, not just high performing leaders. We need to learn the skills they use to delegate to realise the same benefits.
The awkwardness is a sign you are learning and making progress. Every experiment brings you closer to the kind of leverage that transforms your workflow. You do not need a perfect system; you just need to start. Share your experiences, keep experimenting, and watch as the payoff grows.
More articles
Kill Your Prompts: Build Agents That Actually Work
Every technical team I talk to faces the same painful truth about AI agents. They build something that works brilliantly in their demo, showing it off to stakeholders who nod approvingly. Then they deploy it to production and watch it break in ways they never imagined. The carefully crafted prompts they spent weeks perfecting become a maintenance nightmare.
Recently I showed a (virtual) room full of technical leaders how to kill their prompts entirely. I do not mean improve them or optimise them. I mean kill them completely and build something better.
Read more →Master Prompt Stacking: The Secret to Making AI Code Like You Do
After months of fighting with inconsistent AI coding results, I discovered something that changed how I work with tools like Cursor. The problem was not my prompts. The problem was that I had no idea what else was being fed into the AI alongside my requests.
During a recent webinar, I walked through this discovery with a group of engineers who were facing the same frustrations. What we uncovered was both obvious and surprising: AI coding tools are far more complex than they appear on the surface.
Read more →The Huge List of AI Tools: What's Actually Worth Using in June 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 June 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 more →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.
Read more →Building 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