How I'm Using AI to Build My Business (And What It's Doing to My Brain)
AI Is Making Me Better. It's Also Making Me a Little Crazy.
After spending the past three weeks going deeper than I ever expected I’d go on MLS policy, private listings, and the kind of housing market the public deserves, it’s time to come up for a bit of air.
These issues matter to me, and I’ll continue to engage on them as new developments arise. But not today.
Today’s post is about AI.
And I don’t blame you one bit if you groaned upon reading that sentence. I, too, am feeling the fatigue from everything being about AI. But when I asked Claude what I should write about, this is what it suggested, so . . .
I kid. I kid.
For better or worse, this post originated entirely in my own brain. Though, full disclosure, some of my thinking and edits were sharpened by the robots. It is what it is.
Last week, my conversation with Rajeev Sajja went live on his podcast Real Estate AI Flash. You can catch the full episode here. Rajeev and I recorded in mid-February and, while I believe the main thesis of treating AI as your “second brain” is an evergreen idea, it’s incredible to me how much my own usage of AI has changed in just over a month.
Here, in the spirit of collaboration over competition, I want to pull back the curtain on how I’m implementing AI into my everyday workflows to build my businesses. I trust there will be some genuinely useful takeaways for someone somewhere from that alone. But lest this post be all sunshine and rainbows about the brilliant future AI is about to usher in, I want to share some of the darkness I’ve felt pulling on the corners of my mind the deeper I’ve gotten into integrating AI into how I work.
Being spring and all, let’s begin with the sunshine and rainbows.
On the podcast, I shared with Rajeev that for the past year or so, I’ve been orienting toward ChatGPT and Claude as my ‘sparring partners.’ Insofar as I possess the self-awareness to say this accurately, I do not outsource my thinking — hence the Realtor Gone Rogue moniker. Like all of us, I have my influences and proclivities to certain ideas over others, but I find arriving at my own conclusions to be a profound joy in life. Naturally then, I’m not particularly prone to letting robots tell me what to think.
What I am prone to is arguing. And AI, it turns out, is a remarkably patient and fair debate partner.
When I’m working on a piece for this publication, I don’t ask Claude to write it for me. I share drafts and snippets of my thought, then ask it to push back. Hard. I want the best counterargument to my position, not a pat on the head. If I’m wrong about something, I’d rather find out in a conversation with an AI at 10 PM than in a public comment section the next morning. When I was developing my argument about the MLS data integrity debate a few weeks ago, I had Claude argue Compass’s position as forcefully as possible. Doing so actually shifted part of my thinking. Regularly, these arguments with AI lead to ‘unlocks’ in my mind that help me ask better questions, explore new angles, and make connections I would have likely missed otherwise.
There’s a significant difference between asking AI to do your homework and using it as a research partner who happens to have read everything. When I need to understand how Washington’s SB 6091 defines “public marketing,” I don’t want a mere summary. I want to interrogate the language, test edge cases, and figure out where the law is ambiguous. AI lets me do that in an hour instead of a week. The questions and thinking are still mine. But the timeline and effort for discovery has been greatly compressed. This is both a blessing and a curse . . . but more on that in a few paragraphs.
On the building side is where AI has blown my mind most in the past month since I recorded with Rajeev. In early March, I rebuilt my entire brokerage website using Claude Code. I didn’t merely ask Claude ‘to make me a website.’ Over the course of two days, I worked alongside Claude by making design and stylistic decisions, giving examples of what I wanted the site to feel and sound like, and marveling as Claude took my text input to write, test, debug, and ultimately publish the code. I was able to build a full featured, completely custom site for just the cost of a $99/month Claude Max subscription. And, best of all, I’m quite proud of it — you’re welcome to check it out.
I’ve used Cowork to research and build out community pages for new construction developments across Clark County, each with localized data and market context that would have taken me weeks to compile manually. I’ve built a CC&R risk assessment tool that analyzes HOA documents and flags issues for my buyer clients. I’ve created a Reddit monitoring system that surfaces conversations where I can add value and surfaces content ideas I can repurpose.
None of these tools existed six months ago. I built all of them within a matter of weeks. And I still don’t know how to write a single line of code.
That’s the sunshine.
Most posts on this topic either conclude in a technocratic praise of the utopian age AI will lead us into or an anti-tech, ‘AI will destroy us all’ luddite-esque rallying-cry.
I’m resisting both. There’s a genuine excitement I have for the ways AI tools have expanded my horizons of possibility. But there are also some dark clouds on that horizon, and I want to speak honestly about them rather than pretend they don’t exist. By naming these things, I expect some of you will relate to what I’m saying. For others of you who are just at the beginning of your journey with AI, hopefully these categories will help you make sense of what you feel as you engage more deeply with these tools.
Despite the popular rhetoric about AI replacing human work, AI has not helped me work less. Quite the opposite in fact. With AI, I find myself working more. Not just doing more, but actually spending more hours of my week working. Significantly more.
When the only constraint on building something is your own imagination and a few hours of conversation with an AI, the bottleneck is . . . you. It used to be that ideas outpaced my capacity to execute. Now execution is nearly frictionless, and I’m the one holding up productivity. My creativity. My energy. My ability to evaluate whether the idea that just popped into my head is actually worth building.
The only felt limitations to what I can build and optimize are my own lack of imagination and my human needs — eating, sleeping, seeing my family, being present with friends. That might sound like a productivity dream. Currently, it doesn’t feel like one. It’s like putting a jet engine on the top of a Honda Civic. While the car could go faster, one must wonder if the chassis is really capable of all that extra potential.
For me, there’s a low lying mental fog accompanying this unprecedented level of personal output. I’ve never experienced anything quite like it. What used to take months now takes days. What took days gets done in hours. All at a fraction of the cost. That compression of time and productivity makes it genuinely hard to keep track of what I’ve accomplished, to remember and fully leverage what I’ve already built, and it produces a kind of mental exhaustion at the end of the day that’s completely new to me. It’s not physical tiredness. It’s not burnout in the traditional sense. It’s more like the feeling you get after making too many decisions in a row — except it never lets up, because there’s always one more thing you could (or, as I tell myself constantly, should) build before bed.
And undergirding all of it is a low-grade anxiety. A sense that there’s some artificial clock ticking. Yes, AI has helped me level up exponentially at virtually no cost. But if that’s what it’s done for me — a small real estate business with no funding and no technical background — what is it doing for my competitors? The ones who are bigger, better-funded, and have actual engineering teams? There is a real FOMO in the AI space. A sense that what I don’t know I don’t know is going to bite me, and that no matter what I build, someone else will build it better and faster in a matter of days. How do you build a moat in this kind of competitive landscape? The only answer I can muster is, you don’t stop building. You keep utilizing AI and maximizing your efficiency. Which means you work more. Which means the fog gets thicker.
That’s the vicious cycle, and to some degree I’m in it.
There’s also a distinction I’ve been sitting with between efficiency and effectiveness. My efficiency has exploded. My effectiveness has not followed the same trajectory. I am certainly more effective with AI than without — but not exponentially so.
Here’s what I mean. Because AI lets me build virtually anything, I experiment constantly. I built a custom CRM with automated workflows and operating rituals that, honestly, probably exceeds what my business actually needs. I spent hours building out another custom site to showcase all new construction in Clark County — a project that, in the end, I’ll never use. I’ve gone down countless research rabbit holes, investigating questions that, before AI, I would have let go. And I would have been fine never diving deeper.
In the old world, friction was a natural filter. The effort required to pursue an idea served as a gut check: is this worth the time? AI removed that friction, and now there’s nothing stopping me from going deep on everything except my own finite capacity as a human being. The result is that I’m attempting more, building more, and learning more than ever before. But I’m also failing more often. Speed to failure is faster, which is good — in a sense. But psychologically, the volume of dead ends adds up, and it’s adding up faster than ever before. I’m still learning how to not let the sense of spinning my wheels settle in when I’m producing more than ever but still unsure how much of it matters.
Unfortunately, there’s no tidy conclusion here. I’m not going to tell you that the key is “balance” or “intentionality” or whatever nicety people tend to use to make a messy reality sound manageable.
What I will say is this: I think a lot of people are feeling some version of what I’ve described and not talking about it, because the dominant narrative around AI right now is that it’s either going to save us or destroy us, and there’s very little space for “it’s making me better and also making me a little crazy.”
But that’s the essential conversation I think we must begin having. Especially for those of us who are leaning heavily upon AI to build modern businesses, we must care for one another by asking not only, “What are you using AI to build?” but also going a step further and questioning, “How is AI affecting your sanity?”
Circling back to my conversation with Rajeev, in the podcast I told him I’m using AI as my second brain. But I think we must all be critical of the impact our second brains are having upon our first one.
Our AI tools are getting better every day.
Let’s make sure we are too.
Wanna connect further?
Tiered & Flat Fee Real Estate Services in SW Washington: The Tartan Team
DIY Homebuyer Resources & Advocacy: DIY Homebuyer Academy
Connect with me on socials (I’m most active on LinkedIn) @NickAufenkamp