Skip to content
Every AI Tool We’re Using & 4 Pillars to Guide You: Operators Build (E7) artwork

Every AI Tool We’re Using & 4 Pillars to Guide You: Operators Build (E7)

Operators

October 5, 2026

“They’re the most capable ever, but not trained on your business.” What pillars should every AI-first brand build, and which one comes first? Matt Bertulli (Founder, Pela Case & Lomi) and Cody Plofker (Advisor & fmr.

Speakers Matt Bertulli, Cody Plofker

TopicsEntrepreneurshipBusiness

Matt Bertulli (0:00)

Okay, we're here, it is episode seven of Operators Build. This is the all things AI episode. I want to start by saying, I think there's like four big things that we've created that I think are like pre-universal, everybody should do. So the first is, you have to think of company brain, right? This is like context, memory, all the stuff that like the models are just bad at.

Cody Plofker (0:26)

Probably the most important part.

Matt Bertulli (0:27)

Number two would be like an operating system. So it's nice to use all these tools. It's great to build software, but you need some way for the humans and the company to interact and visualize things, right? So like we call that Winks OS, agreed? Number two, all right. Number three, all of the software and tools. So I don't know about you, man, but I'm just to believe it. There's a lot of stuff that we shouldn't be building, that we should just buy and we should just use. So this is like the Omnisense, the Richpanels, the Iris, Highbeam, all of the tools that we've opted to buy for the company. Some of these are sponsors, some are not. And the number four, we'll call them agents. So for Winks, we have a researcher, a brief writer, and we probably have other ones, like competitor analyst, yeah. Those are the four big buckets. And I think what we do today, Cody, is go very deep on each of them.

And we even show people, let's share a screen, we'll get into how we're actually constructing the company. All right, let's get into it.

Explain brain to people.

Cody Plofker (1:29)

The brain is your context layer. And so context engineering is one of the most important things. With these agents, with the Claw to Codex, they're the smartest, most capable things or people ever, but they're not trained on your specific business. They have a lot of general knowledge, and what you need to do is equip them with specific knowledge. And I think about it the same way that you would want to educate somebody if they're coming into your business. What's all the context you would want them to know and have access to? What is your mission? What is your vision? What are your past tests? What are your products? What are your COGs? What are your SKUs? What are your reviews? What are your customer insights? You just want that dumped into one place. And then these agents and any tool, AI tools that you're using, will then be able to access that and have knowledge of it. So you won't have to prompt it every time and say, hey, this is my company. This is what I want you to do. It'll just know what to do if you have the right setup.

Matt Bertulli (2:22)

And the thing with this is important is it's updating all the time. So you don't set something like this up once.

It's a living, breathing, that's why you call it a company brain. It's like you have to have it.

Cody Plofker (2:32)

Oh, yes.

Matt Bertulli (2:33)

Let's go to number two then, the operating system. So one of the first things you did when we started on Winks was you set up Winks OS. What was the thinking? Why is that important?

Cody Plofker (2:41)

I'll say this is one that's probably not the most important. Things just live in so many different places.

We would be getting ready for a launch and we would need cogs from this place, we would need shade descriptions from this place, and copy here, and it just takes so long. I think the big theme is just everything in one place, whether that's for a human or for an agent, have everything in one place that it can access.

Matt Bertulli (3:02)

See, I disagree. I think that it is actually really important. Because if you've ever used one of these LLMs, like to people watching, listening, they aren't very fast. So if I had to say like, can you show me the last 10 ads that we've made today?

That's going to take a long time. Whereas if I go into Winks OS, I can just click and see them. They're there. So there are certain things, like you should not be using all of this compute power from these models to do. So I still think it's more important. Okay, let's move on to software. I think there's a lot of people out there that just think that it's stupid to pay for software in general. Now, where do you stand on this subject? Because we're using a bunch of software.

61 more minutes of transcript below

Thousands of transcripts fetched by people building searchable podcast archives

Fetch the whole transcript

The demo key returns a sample episode in full, no card needed:

request
curl -H "x-api-key: pt_demo" \
  https://spoken.md/transcripts/1000651996090

Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.

From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.

Using your own key:

request
curl -H "x-api-key: YOUR_KEY" \
  https://spoken.md/transcripts/1000793252077