I’ve been using AI for a while, but this summer I started thinking about it differently.
Instead of asking, “What can AI do for me?” I started asking:
“What can I hand over to AI?”
That sounds like a small difference, but it has completely changed the way I think about my workflow.
One of the biggest things I’ve learned recently is how to use an AI Chief of Staff—one central AI assistant that can coordinate with other agents, gather information from different systems, and bring the important updates back to me.
The biggest benefit so far?
Speed.
The Idea That Changed My Thinking
I had heard people talk about creating an AI Chief of Staff before.
At the time, I didn’t think I needed one.
That changed after attending Invest Fest.
While I was there, I heard Alicia Lyttle discuss the idea of having a Chief of Staff that could manage other AI agents. Instead of you constantly checking every agent yourself, those agents could report to the Chief of Staff. Then the Chief of Staff reports to you.
That got my attention.
I started thinking about all the different systems I use and all the time I spend moving between them.
- GitHub.
- CRMs.
- Email.
- Calendar.
- Content creation.
- Task management.
- App development.
- Quality testing.
Each system might make sense on its own. The problem comes when you become the person responsible for constantly moving information between all of them.
That’s where I started seeing the potential.
Stop Checking Everything Yourself
Think about a normal workday.
- You want to know the status of a project, so you log into one system.
- Then you need information about development, so you check GitHub.
- Then you check email.
- Then your calendar.
- Then your tasks.
- Then another application.
Before you know it, you’ve spent a chunk of your day looking for information instead of using information. That’s the problem I wanted to solve. My goal became creating one place where I could ask:
What’s going on?
And get the information I need. Instead of me going to every system, I want the systems to report back to my AI Chief of Staff. Then my Chief of Staff reports to me.
The Biggest Difference Has Been Speed
This is where I started seeing the real benefit.
I’ve been able to access information from my CRM and GitHub repositories quickly without having to log into each portal and search around for updates. I can ask for information and get it. That changes the workflow. It also changes what I spend my time doing.
I don’t necessarily need to spend my time finding every status update myself. I need enough information to understand what is happening, make decisions, and decide what needs my attention next.
That’s a much better use of my time.
And I think that’s where AI becomes much more interesting for entrepreneurs and creators. It isn’t just about generating something faster. It’s about reducing the amount of routine work that requires your direct attention.
I Still Had to Build the Bridge
There is an important part of this story that I don’t want to leave out.
My AI Chief of Staff is housed in ChatGPT, but simply creating an assistant and calling it a “Chief of Staff” wasn’t enough. It needed access to information outside of ChatGPT.
So I had to build an API that acts as the bridge between my Chief of Staff and the systems I want it to communicate with. That bridge allows it to connect to places such as GitHub and my CRM. And I’m continuing to add more connections.
That technical layer matters because an AI assistant can only help with the information and actions it has access to. The more useful question becomes:
What information does my assistant need access to in order to actually help me run things?
That’s a very different conversation from simply asking AI to write an email or create a post.
Where I Want to Take This Next
I’m still building this system.
There are several areas I eventually want to hand over—or at least partially hand over—to my AI workflow.
- Content creation is one.
- Status updates are another.
- I also want better connections with the Google tools I use for tasks, email, and calendar management.
- And then there’s development.
As I work on getting more apps built, tested, and running, I want AI agents helping monitor development progress, quality, and testing.
The idea isn’t that AI makes every decision for me. I still make the decisions. What I want to eliminate is the unnecessary work required just to get enough information to make those decisions.
There’s a difference.
Think Like a Manager, Not Just an AI User
This may be the biggest lesson I’ve taken from the experience. A lot of us are still approaching AI as individual tools.
- Need an image? Open an AI image tool.
- Need some writing? Open ChatGPT.
- Need to check development? Open GitHub.
- Need a status update? Go find it.
But what happens when you start thinking about AI more like a team?
One agent has a responsibility. Another agent has a different responsibility. They gather information and perform specific jobs.
Then instead of managing every interaction yourself, you create a central assistant that helps coordinate everything. At that point, your role starts changing. You’re not just prompting AI anymore. You’re managing a system. That may be one of the most important shifts for people trying to become more productive with AI.
Start With the Bottleneck
You don’t necessarily need to build the system I’m building. You also don’t need ten AI agents tomorrow. Start smaller. Look at your week and ask yourself:
What am I repeatedly doing that takes time but doesn’t really require me to do it personally?
- Maybe you’re constantly searching for project updates.
- Maybe you’re manually organizing tasks.
- Maybe you’re checking multiple calendars.
- Maybe you’re spending hours creating variations of content.
- Maybe important information is scattered across too many platforms.
Find the bottleneck first. Then ask whether AI can help collect, organize, summarize, draft, monitor, or route some of that work. That’s where I think the real opportunity is.
AI Should Give You Back Time
I’m still learning. I’m still connecting systems. I’m still figuring out what should be automated, what should be delegated to an AI agent, and what still needs my direct attention. But this summer has changed the way I think about AI. For me, the goal isn’t to use more AI just for the sake of saying I’m using AI.
The goal is to build a system that helps me move faster, stay organized, and spend more of my time on the work where I actually need to be involved.
If you’re using AI already, maybe the next step isn’t finding another AI tool.
Maybe the next step is figuring out how the tools you already use can start working together. Maybe instead of asking AI to do one more task, it’s time to ask:
What part of my workflow can I stop managing manually?
That’s the question I’m working on now. And I think it’s a question worth asking.