“Nipsey Hussle is from Crenshaw. You, Nifemi Aluko, you’re from Ikeja Lagos. So go find your own lane. The marathon continues.”
I almost spilled my coffee laughing when my brand consultant gave me that very direct feedback. Then gave me a pep talk. I worked through my recent rebrand. We had aligned everything. I had brainstormed my offer stack and linked my newsletter to my long-term goals. Analyzed the top 20% of my newsletter to see what has performed best in the last 3 years. I even analyzed the feedback that some of the readers have given. It was all in a document.
After rigor and analysis. I landed at a place where I was happy with my consultation.
I had a name. A bio. A strategy. Good to go.
I thought I’d sleep on it before locking it in.
Then the thing that happens to most entrepreneurs happened. I woke up with a brand new idea. I thought “Ahh, I forgot this. Let me do a comparison of my brand direction with a brand I really like.”
I asked my consultant, “So I love “The marathon continues” brand, how does my brand compare to what we created yesterday.”
My consultant paused. The recommendation from yesterday’s session was literally “begin executing.” But here I was with another question. He responded with how the brand is aligned and why they are separate.
Then he hit me with the hard truth:
“Go find your own lane. Now go execute.”
My consultant’s name is Claude.
It’s the AI chatbot and assistant from Anthropic. It has become my go-to business companion. But like a co-worker that you learn to work with, I’m still learning how to get the best results from it.
This made me realize, AI can become your difficult coworker soon. If that’s the case, how do you learn to work with it like one of your colleagues?
Your AI coworker doesn’t read minds
The more I use AI the more I realize that I just have to learn how to interact and collaborate with it.
The same way you have to learn to “deal with a co-worker.” Putting your hand up, stomping your feet, saying you’re not going to work with AI, is not going to help you. The same way saying you’re not going to work with a new “difficult teammate” doesn’t help your team.
Here’s what I’ve noticed: when AI gives me garbage output, I’m quick to blame the ‘stupid AI’ or complain about how ‘overhyped’ these tools are. But when it produces something brilliant, I’m ready to take credit for my ‘clever prompting.’ The more I work with AI tools, especially building actual products, I’ve learned that most of what I call “bad AI response” are actually collaboration failures.
The AI didn’t misunderstand me. I just wasn’t clear about what I wanted.
It didn’t give me bad code. I just didn’t provide enough context.
Most of the frustration comes from expecting it to read my mind instead of treating it like what it is: a teammate that needs clear communication.
My challenge with working with AI has been the switching cost of my attention. Switching from the task of prompting it to give me an outcome, waiting for that outcome, and then realizing the outcome is not what I want. And poor communication begets poor results that beget even worse communication before you know it, your tailbone is screeching down the muddied rabbit hole of frustration.
For instance I was building a tool this weekend: it seemed easy at first, but it started taking up time.
That’s because I didn’t want to pass PII (personal identifiable information e.g someone’s email and name) through the AI model. I had to figure out a workflow to achieve this along with all the different tools and integrations. Then set up an evaluation process to be able to improve the application. As I worked on this, I realized you can keep prompting and get lost in the workflow unless you have a systematic way to build things in a modular way and approach it step by step.
By the end of the day I had something I could easily test and send to a few people for feedback. If you want to try it out here it is.
AI teamwork is an interpersonal skill
My realization (similar to all the other apps I’ve built with AI) is that this stuff is a lot faster than what I could dream of building five years ago but it is also not as fast as you think unless you have a mental model of how to build already.
You’ll get frustrated. You will want to yell at it, only if it could listen to your tone. You’ll go from moments of awe (this is so amazing) to valleys of disappointment (this stuff is all hype). You will be tempted to point fingers at it and go yell from the LinkedIn pulpit about how “underdeveloped” these over-hyped AI assistants are.
But at the end of the day, you get to interact with it. You begin to learn your own style. And when you get either frustrated or excited, you learn to deal with your expectations and manage your emotions.
In essence you’re building “self-awareness” to function in the automation era.
Same thing like when you can build a great company with a team, if you don’t know how to get along with your teammates, that company is going nowhere.
It’s more about learning how you are going to work with this teammate. The sooner you start, the better.
Learn to work with AI. It will soon become your difficult co-worker. Unless you do begin the interpersonal dynamics work today.
Documentation as a practice:
In an era of constant content creation, documentation and annotation is even more important.
I currently have structured documents where I write responses from AI queries that I then read fully before I start prompting again. Without this, you run the risk of doom-prompting (the AI chatbot version of scrolling endlessly on social media, where you keep prompting for prompting sake). You’ll be like the colleague that keeps talking about his work at the water cooler but never gets to it.
When you document what you build, you give yourself a chance to make progress. It gives you the ability to keep track of what you’re building.
For instance, with the recent tool I was building. Once I was done iterating, getting advice, and building out the workflow, I asked the AI assistant to summarize everything we had worked on, a visual of the entire workflow, and steps to test and iterate on the next steps.
I saved that in a document.
When next I start on the project again, I will import the entire document to give a snapshot of where I’m at, and I might even ask it to give me a structured prompt to use to achieve my next step.
This type of systematic approach avoids time-wasting that’s not good for your project and personal health.
Interpersonal dynamics with AI
My AI consultant told me to find my own lane and go execute.
It turns out, that applies to more than just branding. It’s about finding your own way to work with these tools too. Learning to collaborate with AI is like any other practice on the ground: messy, iterative, and requiring self-awareness. The documentation habit is just one experiment. The real work is building your own interpersonal dynamics for the automation era.
Start now. Your future difficult coworker is already here.
Yours truly,
Nifemi
P.S. What’s your biggest AI collaboration challenge? Hit reply or comment below – I read every response and often feature reader experiments in future editions.
