Hey!

Chris here. Welcome to Blueprint—the newsletter to help you build a winning engineering team.

Lately, when a founder comes to me with fresh capital and a clean slate, the conversation often starts the same way:

"Which AI tools should we build around?" "Which platforms are winning?"

It's the natural place to land. It's also where things go sideways.

Let me show you where I'd start. 👇

📒 DEEP DIVE

How to Build a Company for the Agentic Era

The 6-step playbook for founders who want to build around AI the right way. 

Step 1: Start at intake

The first question should be about process: What is the intake of your business?

Start with whatever it is you're trying to get—a lead, a customer coming in for support, a service request—then walk the path all the way through to delivery, completion, or implementation. Whatever it is you ultimately get paid for.

Many founders skip this step because they assume they already know how their business works. Most don't—or at least not well enough to see what's actually possible now.

Step 2: Map what can go on rails

Once you've walked the full chain, you start to see the opportunities. And they're bigger than you think.

Almost no one will know what's possible along that chain. This is exactly why having a guide matters right now. The conversations I've had with companies at the intake level alone have been eye-opening. Just the way you get business can be put on rails. And that's before you've touched anything else in the chain.

Walk the same logic all the way through delivery. At every point, many processes that look human-centric don't have to be that way anymore.

Step 3: Accept that the org chart changes

This is where it gets uncomfortable for a lot of founders.

When you redesign work around agentic flows, many of the roles that exist because of human handoffs stop making sense the same way. The reason certain roles exist was often about passing work from one person to the next. When that changes, the roles change.

The honest version of this: we don't even have words yet for what some of those new roles will be. I'm absolutely certain that within the next 5 years, there will be titles where, if you heard them today, you wouldn't recognize them. From C-suite roles all the way down to boots on the ground.

Many founders are still drawing the old org chart. That map doesn't match the territory anymore.

Step 4: Kill the ceremonies

Software is the sharpest example of how inherited practices break down in this new world, so I'll use it—and I'll say the thing that makes some people angry.

For years, there was a battle: waterfall vs. Agile. Agile won, and the ceremonies followed: story pointing features, writing PRDs, sprint planning, standups, and 2-week cycles.

But none of that matters anymore. I'll go further and say I don't think most of it mattered before, either.

Some companies at a certain scale genuinely needed it—I'll give them that. But spending weeks in planning ceremonies to break down work that an AI agent can prototype in an afternoon is a ritual that outlived its purpose.

Step 5: Make product building a creative conversation

The features you should build are not the best when they’re only out of your own head—even when supported by customer research. Brainstorm them with AI.

Your product will be better if this is a creative conversation, where you're the creative one, prying, asking questions, poking AI agents to help you prototype quickly so you can see what you're actually building before you've committed to anything.

The new operating model: have an idea → make a prototype in minutes → iterate faster than ever.

Step 6: Let taste and judgment drive the output

There's a narrative out there that all of this is vibe-coded slop. That anything AI-touched is suspect.

But the truth is that there was crappy, horribly insecure code well before AI came on the scene. That was a human output. And that's still the case.

If you produce low-quality, insecure code today, it's because you guided the process terribly, didn't have good context around it, and didn't know the sequence of steps you needed to follow. You guided it that way.

If you know what good looks like, you're not going to stop on those iterative cycles until you have something good. AI is just leverage—you create more of the thing you're already great at making.

What determines quality is whether someone who knows what good looks like is driving the system.

Why You Should Question Everything You Inherited

When you're looking at how to build the company of tomorrow, you have to think about your process from beginning to end and question every single role and every best practice you always believed to be true.

Because some of them aren't true anymore. Some of them changed, and some of them—this is the part that gets overlooked—are more important than they ever were.

And understanding which is which starts with understanding your own process first.

BEFORE YOU GO…

The best practices around agentic companies are still being debated and figured out in real time—and that includes by me.

But that's not a reason to wait. The people who are learning right now will be years ahead of the people who hold off until someone hands them a finished manual.

That's why if I were you, I'd get started today.

Talk soon,

Chris.