Read this if your team won't stop saying "agent"

These days, if I talk to someone about AI, I can almost guarantee they'll say the word "agent" in the first 10 minutes.

But when I poke around at how they're using AI, most of them aren't doing anything remotely agentic.

They get talked about so vaguely that no one feels like it's their job to figure them out. 

Nobody is sitting down to learn the practical stuff: what agents are, how they can help, how to set them up...

Because that gets skipped, agents end up feeling like something technical teams handle. Which also means someone else is deciding how AI shows up in your work and for your team.

But agents are far more approachable than the way everyone talks about them. 

You just have to start with the foundation: actually getting good at the basics first.

My secret for doing that sounds way less impressive than the buzzword.

It's learning to set up projects (or notebooks if you're using Gemini or Copilot). 

You heard me right.

They may sound like a way to organize your chats, but stopping there completely underutilizes them.

Projects are the easiest place to experiment with two things that change results: 

1. Drafting prompt instructions for AI to read at the start of each chat
2. Attaching and adjusting context it can access (files, examples, rules, etc.)
   

When you get the hang of these, projects turn into more than a folder.

Here are some I started with:

• A business advisor that knows my company and goals, for strategy questions
• An automation expert I taught with the n8n docs, for building workflows
• An image prompt drafter with my brand style, for consistent visuals
• A client lens I run work past for the customer's view before shipping

You can create them for an initiative (one place for everything on a topic), a task (a repeating job, done faster), or a specialist (giving feedback from a specific point of view).

Once you start setting up projects, agents stop feeling abstract. 

Agents are built from the same pieces: instructions and context, pointed at a task, with guardrails so they can run on their own.

The bit of set up is what makes AI pay off. It's also where agent work starts. 

I made an interactive lesson to help your team get the most out of projects and notebooks, and I'm excited about this one.

With one link, your team levels up in minutes.

Free team lesson: projects or notebooks

Here's how it works:

1. You select your platform: ChatGPT, Claude, Copilot, or Gemini
2. Send your free lesson link to your team
3. Each person spends 15 mins setting up their own project or notebook
4. You see what they created in your manager dashboard

Create your free team link: chasingnext.com/teams/projects

---

Not part of a team? Find the lessons here.