I Lead a Team of AI Agents. Here Is How You Can Too.

Published: July 15, 2026 • 6 min read

I don't just use AI. I lead a team of them.

Yes, a literal team of AI agents that have evolved since the first member of the team was created in October, 2025. And everything I learned building and leading them is what I am here to teach you, so you can get real hours of your week back.

That chart above is my whole team. 17 AI agents organized into five departments, structured like a real company. Each one owns a specific job. One edits my blog posts, one runs my content automation while I sleep, two coach how I speak in English and French, and one keeps track of every client relationship for me. Those are just four of them.

You will meet the rest of them over time, and as you do I will show you how to build your own versions, so the same jobs get done for you. The most active on my team is Allen Kendrick, who helps me refine my writing without ever replacing my voice.

I describe him as the one who takes a rough sketch and turns it into a polished painting, except the painting still looks exactly like I, the artist, intended.

So when I say "leading AI agents," I mean the human skill of running them like a team you're responsible for, instead of being run by them without ever noticing it happen. That is what I want for you too. You lead them. They do the labor around your thinking.

A real human with a real life is exactly who should be leading AI, not disappearing behind it. That is the whole reason this publication exists, and the exact skill I want to hand you, so you run your tools in a way that gets you real results.

Who I am, and why you might trust me on this

I'm Prisca, an AI Consultant based in Montreal, Canada and I work in both English and French.

The honest story of how I got here is that after I graduated in 2025, I sent hundreds of job applications into the void for about 40 days and got mostly silence back. So I stopped applying and gave myself 50 days to build public proof of work instead.

I experimented heavily with AI, I learned things, I built things, I broke things, and I documented the entire messy process as it happened. I fell in love with writing somewhere in those 50 days, and I have not stopped since.

That period turned into 160+ blog posts written in a span of about 8 months, which tells you something about what happens when I decide to do a thing.

Since then I've shipped custom AI systems for clients, from MCP servers to content intelligence tools. I build systems to eliminate repetitive patterns in people's workflows to help reclaim hours, cut the busywork, and finally see a return on the tools you are already paying for.

I am not writing to you from theory. I'm writing from the middle of the actual work.

I also live with a chronic illness. My stomach has hurt every day for more than 560 days, and I have built all of this while in pain. I write about that here too, which brings me to the part most publications would hide.

This publication is two things, on purpose

You're going to get the builder and the human, and I'm not going to keep them in separate rooms.

Some weeks I'll show you exactly how I work with AI. How I build agents, how I save money on tokens, how I lead a team that lives inside my computer. They would be practical, specific, and useful things that you can take and use on the same day.

Other weeks I'll write about the rest of it. Living and building with a chronic illness. Learning French as an adult and being humbled by it. Building my storytelling and public speaking skills in public. Figuring out who I am when I refuse to be just one thing.

Those posts belong right next to the technical ones for the reason I opened with. To me, the building and the staying human are not two jobs. They are the same one.

If that combination feels a little unusual, good.

I've never been able to be just one thing, and I stopped apologizing for it.

My promise to you

My role here is that of a teacher, and I hold myself to one standard above all others. If something I write does not make sense to you, the simple fact is that I failed to teach it well, and I'll try again.

Everything I share here is proof of work. I show you the process, not just the polished result, including the parts where I broke something and had to figure out why. When I recommend a way of working with AI, it's because I do it myself, not because it sounds good in a post.

And I will always, always help you use AI in a way that makes you more yourself, never less.

Where to Begin

If you're new here, don't try to read everything. Start with whichever version of me you came for.

If you want to actually understand AI, without the jargon:

"Tokenomics for Humans: Every Word Costs Money". The clearest piece I have written on what a token actually is and why every word you send has a price, no jargon and no pretending you already know the basics.
"Tokenomics for Humans: Series Launch". I read a 28-page Deloitte report so you do not have to, and this opens a 12-part series that explains how AI economics really work, one readable piece at a time.

If you want to meet my AI team, and learn to build your own:

"I Let My Blog Editor Introduce Himself. He Said 'It's About Time.'". Meet Allen Kendrick in his own words, the agent I lead the most.
"Claude God Tip #4: Create Your Own Employees". This is the useful companion to the introductions, because it shows you how to build your own agents instead of just admiring mine.
"Igor Jarvis is Now Autonomous. Here's What a Week Looks Like.". This is what it looks like when an agent runs on its own schedule and does real work while I sleep.
"Who Needs a CRM? Thomas Raynor Knows More About My Clients Than I Do". Meet Thomas Raynor, the agent who keeps track of every client relationship for me.
"I Told My AI Coaches to Stop Protecting Me From Mistakes". Meet Oprah Winfrey and Tiana Picker, my English and French speaking coaches, in the moment they evolved from polishing my scripts to catching how I actually speak in real time.

If you're already paying for AI and want to actually get your money's worth:

"Why You're Overpaying for Claude Max". Most people paying $100 to $200 a month would be perfectly fine on the $20 plan, and this walks you through the token math that proves it, so you stop paying for capacity you never touch.
"Claude Pro vs Max: The Complete Decision Guide for 2026". If you have ever stared at the pricing page unsure which plan you actually need, this settles it with a simple decision tree, not a sales pitch.

If you want the human behind the work:

"My Stomach Has Hurt for 565 Days. I Finally Found Peace That Doesn't Wait for the Pain to Stop." ([Substack link pending]). This is a piece I wrote about finding peace while living with a chronic illness, and the clearest picture of why I write about my life right next to my AI work.

Proof, not promises

I could tell you all of this works. It's more honest to let the people who used it tell you instead.

"The agent has helped me go from spending a day planning out ideas to doing the work in a few hours." — Theo, after using my Blog Refiner agent and LinkedIn post builder

A day of planning turned into a few hours of doing. That is the whole point of leading these agents well, and it is the exact return I want you to feel.

Melissa Maughn thanked me publicly on LinkedIn for a Claude prompt I shared, the one she used to create the image for one of her own posts. She took something I gave away for free and made her own work with it, which is precisely what I hope you do with everything here.

[Image placeholder: screenshot of Melissa Maughn's public LinkedIn comment thanking me for the Claude image prompt. Insert on the live site here.]

And then there was the reader who told me why they follow the work at all.

"I was truly looking for the source. The architect. The builder. The mastermind. Not the messenger." — a reader, on why they follow my work

I keep that one close, because it names the thing I am trying to be. Not someone repeating what the tools can do, but someone in the middle of actually building with them.

Most people are paying for AI tools and barely getting their money's worth out of them. That is the gap I care about closing. I want to help you turn AI from hype into real results, the kind you can measure in hours you get back and busywork that stops eating your week.

I'll show you how to do that without disappearing behind the tools, because the goal is not to become like everyone else who owns a chatbot. The goal is to get real returns while fully embracing all that makes you you.

Subscribe and I will help you get more back from the AI tools you already pay for, one usable post at a time.

[Subscribe button placeholder: wire the live Substack subscribe button here.]

Thank you for being here.

Prisca

I Lead a Team of AI Agents. Here Is How You Can Too.