Screensharing top takes in AI/startups

**Greg Isenberg** (0:00)
And we are live.
Welcome to SIP Live, where we live stream, answer your questions, look at the timeline, give you our takes. I'm your host, Greg Isenberg, and I'm here with my co-host, Jonathan Courtney. He's a shrimp millionaire from Berlin, originally from Ireland. Welcome to the show, sir. We're gonna get right into it. Yesterday, I posted, what's the one thing you'd ask me and my fellow shrimp millionaire over here, if you had me on the phone? And we got 139 responses. I'm gonna go through three or four of them with you, Jonathan, right now. And then at the end of the stream, we're gonna go through a few more. And in the middle, we're just gonna go and talk timeline takes.
This is coming from the X Chat. We got M Pesci says, when you begin the process of making a company AI native, where do you begin? Is it departmental change, or department by department you're changing things, or is it company-wide? Where do you start? And what are first steps? For me, I mean, I could spend four hours just talking about this, but the number one thing that I think is important is to make your company legible.
So what does that mean? That means that if you're using an LLM, you need to have the LLM be able to talk to your company. And that means your meeting notes need to be legible from the LLM. That means that your SOPs need to be legible. So basically going through everything in your company and then putting it so that it could be seen by whatever, if you're using Cloud Code, if you're using the Codex and the OpenAI ecosystem, the Google ecosystem, whatever it is, just make sure that it's being read. And then you start thinking about, that's step one. And then step two is going role by role and then basically saying, okay, this person, what are the jobs to be done for that person? So if it's a social media marketer, they need to do research around what content they should be creating. They need to actually be creating the content. They need to be scheduling the content. They need to be looking at the analytics. And then they need to be taking that feedback from the analytics and then bringing it back to the front so that they could say, we want more of this content or less of that content. So what you end up doing is you build loops, you build AI employees, essentially, that end up doing some of those things. And then some of those things, you're going to obviously remain human. Those are the two biggest things you can do if you want to go AI native.

**Jonathan Courtney** (2:54)
Can I just add one thing to that? What we did, so we, I had already three businesses before all the AI stuff kicked off. So we had to figure out, how do we go from where we are now to adding these AI tools in to a system that's already kind of working. And also, not everybody in the company was excited about integrating AI tools. That's another thing. I mean, if you're asking this question from the perspective of a brand new company, it's a little easier. But I was building on top of what already existed. And so, the first phase, honestly, was just giving everybody the pro versions of ChachiBT and Claude, letting people mess around with it for a while without any rules, because I know that there was these companies where you have to use this many tokens, or you have to automate all these things. I just let people use this stuff for a while, and then they saw me using things, getting excited about things. Then they would show me things, getting excited about things. We just started to see that people individually found little shortcuts and little fun ways to use the tools.
Then the only thing that I did sort of mandate was that the admin part of AJN Smart was the first phase where creating proposals, pulling together paperwork for tax-related things.
Basically, repetitive tasks, that's something where I wanted us to start using these tools so that we could free up time to do other things like optimizing our revenue versus just trying to do all the paperwork. And so for us, it was just a matter of slowly letting people get excited and using this stuff, and then starting with one department, in our case, Admin Bookkeeping Paperwork Legal. And that's where the AI tools started to come in. And now it's sort of spread throughout the entire company. But yeah, we didn't have like an overall top-down mandate that everyone has to use the tools.