I get this question a lot, and it usually comes out a little embarrassed, like the person thinks it's a dumb question. It isn't. Here's the shape it takes: "Tim, everybody's telling me to build some custom AI thing for my business. But I already pay for ChatGPT. Do I actually need to build anything, or am I fine?"
My honest answer is: most of the time, for most people, you're fine. And I say that as somebody who has spent a lot of nights building his own stack. So let me walk you through how I actually think about this, in real dollars and real hours, because that's the only way to make the call. My thinking is that build-versus-buy isn't a philosophy question. It's an arithmetic question. Let's do the arithmetic.
First, get the words straight
Before we compare anything, I want to make sure we're comparing the right two things, because people mush them together.
On one side you've got a general assistant. ChatGPT, Claude, Gemini, the ones you talk to in a chat box. You open it, you ask, it answers. It's broadly smart about almost everything and it costs you something like twenty bucks a month. That's "buy," and honestly it's barely even buy, it's more like rent.
On the other side you've got a custom agent. That's a system you set up to take a goal, break it into steps, use your tools, and carry a task through on its own without you babysitting each move. That's a different animal. That's not "answer my question," that's "go do the follow-up on every new lead for the next thirty days." If the difference there is fuzzy, I wrote a whole piece on it, because it matters: AI agents versus automation, the real difference.
Here's the thing to hold onto. The chat assistant is a tool you pick up. The custom agent is a worker you set up once and let run. You reach for the second one when you've got the same repetitive job happening over and over, not when you just need a smart answer now and then.
The real math on "just use ChatGPT"
Let's price out the simple path first, because most folks skip right past it and they shouldn't.
Twenty dollars a month gets you a genuinely capable assistant. For that money it'll draft your seller letters, rewrite your listing descriptions, summarize a long inspection report, help you think through a tricky negotiation, and answer the same questions a paid consultant would charge you a couple hundred bucks an hour to answer. If you're a solo operator doing a handful of deals, that one subscription probably covers eighty percent of what you actually need AI to do. Eighty percent, for twenty bucks. That's not a starter plan you apologize for. That's the plan.
The catch, and there's always a catch, is that it only works when you're at the wheel. You have to open it. You have to ask. You have to paste the thing in and copy the answer out. It doesn't watch your inbox. It doesn't wake up at 6 a.m. when a lead comes in and start the follow-up. It sits there until you go get it. For a lot of the work a solo operator does, that's completely fine, because you were going to sit down and do that work anyway. The assistant just makes the hour you were already spending into twenty minutes.
So the first honest question isn't "should I build?" It's "am I actually maxing out the twenty-dollar tool?" Most people aren't. If that's you, don't build anything yet. Go squeeze that subscription until it squeaks. You'll learn what AI is genuinely good at in your business, and that education is the thing that tells you what, if anything, is worth building later.
When building actually earns its keep
Now, when does it flip? When does it make sense to build or buy a custom agent instead of driving the chat box yourself?
It flips when you've got a job that is repetitive, that has a real cost when it's late, and that's eating hours you can't get back. The clearest one in my world is seller follow-up. A lead comes in and sits two days in the inbox because you were out running a walkthrough, and that lead is a job that walked out the door. That's not a "help me write a message" problem. That's a "somebody or something needs to respond in the first five minutes, every time, whether I'm there or not" problem. That's an agent's job, and I've written about what AI agents actually do for real estate investors if you want the concrete version.
So here's the arithmetic. Say the repetitive job is costing you five hours a week. Five hours you'd rather spend in front of sellers. Building or buying an agent to handle it might run you anywhere from a thirty-to-a-hundred-dollar-a-month tool up into real money, easily a few thousand dollars, plus real setup time if you go custom. The question is dead simple: is getting those five hours back, plus catching the leads you were dropping, worth more than what the agent costs you in dollars and in the time to stand it up? Sometimes the answer is a clear yes. Plenty of times, at the size most folks are actually operating, the answer is "not yet."
I want to be straight with you about the build side, because I live on it. I built my own stack. I did it because I run a lot of things at once and I genuinely enjoy the engineering, and because I wanted to understand the machine from the inside so I could teach it. But building is not free just because you're not paying a vendor. You pay in your own hours, and your hours are the most expensive thing you own. Every night I spent wiring something together was a night I wasn't in front of a seller. For me that trade made sense. For most solo operators, most of the time, it doesn't, and I'd be doing you dirty to pretend otherwise.
How I'd actually decide
If you handed me your business today and asked what to do, here's the ladder I'd walk you up, in order.
Start on the twenty-dollar assistant and use it hard for a month. Learn where it saves you real time. Then look for the one repetitive, time-sensitive job that keeps costing you money when you're not fast enough, and there's usually exactly one that's the worst offender. For that one job, buy the simplest ready-made tool that does it before you ever think about building. Building your own is the last rung, not the first, and you only climb to it when you've got a need specific enough that nothing off the shelf fits it, and enough volume to pay back the hours you'll sink in.
That order matters because it keeps you from the most common mistake I see, which is spending three weekends building a custom agent to solve a problem a twenty-dollar subscription and ten minutes of your attention would've solved. If you're just getting your footing in the business at all, honestly, put the AI question down for a minute and go get solid on the fundamentals first, the same way I'd point a brand-new investor at how to become a real estate investor before anything else. The tools serve the operator. Never the other way around. And if you want the deeper version of the timing question, I laid it out in when a solo investor should actually start automating.
At the end of the day, "build your own AI or just use ChatGPT" almost always answers itself once you put real numbers on it. Buy first. Build last. And whatever you run, make sure it's giving you back more time with your people than it's taking from you. Those are my people, the ones who have to struggle to put it together, and I don't want any of them burning a month building a machine they didn't need. That's the goal.
Disclaimer: This post is for informational and educational purposes only and is not financial, legal, or investment advice. Real estate carries risk, and individual results will vary depending on your market, your resources, and your effort. Do your own due diligence and consult a qualified professional before making any decisions.
