Why I Use AI to Run My Businesses (And You Should Too)

Key Points

AI has shifted from “interesting tool” to “operating system” for how I run my businesses — it’s the single biggest competitive advantage available to small operators right now
The payoff isn’t theoretical: I save 15-20 hours per week, eliminate entire categories of repetitive work, and access capabilities that would cost $50K+ to hire
The real opportunity for solo operators and small teams isn’t in replacing people — it’s in gaining the leverage of a much larger team without the overhead

Two years ago, I was skeptical of AI’s actual utility for business. Not because I didn’t understand the technology — I did — but because most of the hype felt detached from reality. The demos were impressive. The benchmarks were flashy. But I couldn’t see how it actually moved the needle on revenue, product, or team capacity in the businesses I was running.

I was wrong.

Today, AI doesn’t just power my businesses. It is foundational to how I operate. I use it for strategy, content, customer communications, operations, code review, hiring decisions, and about fifteen other critical functions. Not as an experiment or a side project, but as the core of my workflow. And I’ve watched the impact compound month after month.

This is the piece I’ve been working toward across all my AI writing — the culmination of everything I’ve learned about using artificial intelligence as a serious operational tool. Not the breathless startup-world vision of AI, but the unglamorous reality of what works for someone actually running multiple businesses.

How I Got Here: The Two-Year Journey

My relationship with AI shifted in three distinct phases.

Phase one: Skepticism (2022-2023). I followed the releases. I tested ChatGPT when it launched. But I was building Rotate, running The Points Party, consulting, and shipping projects — I didn’t have time to figure out experimental tools. When people said “AI will change everything,” I mentally filed it next to “blockchain will revolutionize finance.” Interesting. Probably overhyped.

The turning point came in late 2023. I was writing a post about operations systems, and I asked Claude to help me structure the argument. Not write it for me — help me think through it. The quality of the thinking was immediately different from anything I’d experienced with ChatGPT. It felt like having a genuinely smart advisor in the room. That cracked something open.

Phase two: Experimentation (late 2023-mid 2024). I started using Claude for everything I could think of. Writing. Strategy. Debugging code. Brainstorming email sequences. And the consistency was striking. Unlike the brittle, often-wrong previous generation of tools, Claude didn’t hallucinate facts, didn’t confidently give bad advice, and could hold complex context across long conversations.

I realized this wasn’t a productivity hack. This was a leverage point.

I started experimenting with specific workflows. What if I used AI to draft customer emails and just edited them? What if I used it to generate options for marketing copy instead of staring at a blank page myself? What if I used it to audit code before submitting for review? Nothing revolutionary — just thoughtful integration into existing processes.

Phase three: Integration (mid 2024-now). The shift from “using AI sometimes” to “AI is embedded in everything” was gradual but complete. I stopped thinking of Claude as a tool I use and started thinking of it as a team member with specific responsibilities. A fast, always-available, inexhaustibly patient team member who never sleeps and costs less than a coffee a month.

By early 2025, I realized I wasn’t running my businesses the same way anymore. The operating system had fundamentally changed.

The Daily Reality: What a Day Actually Looks Like

Let me walk you through what my day looks like now, because abstractions don’t matter — execution does.

I wake up, check my email, and three customer requests are waiting. Historically, I’d write each response from scratch, context-switching, trying to hit the right tone. Now: I paste the request to Claude, I add context (“Customer is a designer who buys [product], tone should be technical but warm”), and Claude drafts a response in 30 seconds. I read it, edit it for my personal voice (3-5 tweaks usually), and send it. Total time per email: maybe 2 minutes instead of 7.

It sounds small. Across 20 emails a day, it’s an hour saved before my day officially starts.

Next, I’m working on the strategy for one of our products. I’ve got three potential directions, and I’m trying to think through the customer impact of each. Instead of building the mental model alone, I paste in our current strategy doc, our last three months of customer interviews, and our revenue breakdown. I ask Claude to identify which direction best aligns with customer needs and our strengths. I get back a structured analysis that surfaces things I would have eventually found, but in 10 minutes instead of two hours.

By mid-morning, a customer bug comes in. Our engineering team is busy. Instead of waiting or context-switching someone off their work, I open Claude Code — Claude’s IDE integration — paste the issue, and walk through it together. In 45 minutes, I either have a fix I understand deeply or a clear direction for the engineering team. Not every time works perfectly, but the hit rate is high enough that it’s become my default move.

I’m working on a blog post (like this one). I’ve got my outline and my voice guidelines loaded into a conversation. Instead of writing from a blank page, I’m thinking with Claude. I write a rough section, I ask it to strengthen the argument or find the weak spots. It pushes back on vague statements, suggests examples, identifies logical gaps. The final draft is entirely mine — my voice, my perspective, my structure — but the thinking is sharper because I had a smart collaborator.

Somewhere in here, I’m reviewing a hire. We’ve narrowed down to three candidates. I paste in the interview notes and the role description. I ask Claude to identify which signals suggest someone would actually thrive in this role versus just interview well. It’s not the final word — my judgment still matters — but it’s like having another experienced eye on every candidate.

In the afternoon, I’m building a new feature. I’m not a full-time developer, but I code in Astro and Node. I’m working with Claude to think through the architecture. I write the code, run into a bug I don’t immediately understand, and instead of spinning for an hour, I get Claude to help me debug. Not just tell me the answer — actually work through the problem with me. That’s a genuine skill multiplier.

The Points Party newsletter goes out this week. I’ve got a rough draft. Instead of writing from scratch, I ask Claude to take my scattered notes and structure them into the newsletter format we use. I heavily edit it for voice and personality, but the scaffolding is done in minutes.

By day’s end, I’ve had probably 30 AI interactions. Each one saved 15-30 minutes, or saved me from a context switch, or surfaced something I would have missed. That compounds.

The Business Impact: What Actually Changed

The time savings are real, but they’re not the most interesting part. Here’s what actually matters:

I’ve regained 15-20 hours per week. That’s not speculative. I track my time, and the reduction from AI integration is measurable. That’s roughly a day of regained focus per week — enough to ship features that have been waiting, actually write substantial pieces (like this one), and have room for strategic thinking instead of just reaction.

I’ve eliminated entire categories of cognitive busywork. Email drafting. Syntax debugging. Copy iteration. Brainstorming structure for longer documents. Customer onboarding templates. Code review notes. The compounding effect of removing these small drains is that my energy is higher because I’m not spending cognitive capital on stuff that’s not where my judgment matters most.

I’ve gained access to capabilities I can’t hire for. I’m not a designer, but I can generate design options and iterate on them. I’m not a full-time developer, but I can write more substantial code in less time. I’m not a copywriter, but I can produce much better copy faster. In each case, the AI isn’t replacing expertise — it’s giving me enough capability that I don’t need to hire someone just for that specific function.

The quality of strategic thinking has improved. This one is harder to quantify, but it’s real. When I have a sticky business problem, I don’t just think about it in the shower. I walk through it with Claude, get pushed on my assumptions, get exposed to frameworks I might not have considered, and iterate to a better answer. That’s worth more than the time savings.

New projects became feasible. I’ve shipped three side projects in the last year that would have been impossible without AI — not because they’re technically hard, but because I simply don’t have the time to do everything from scratch anymore. With AI handling the scaffolding and the repetitive parts, I can focus on the unique, high-leverage work.

What’s this worth in dollars? If I valued my time at $200/hour — conservative for someone running multiple revenue-generating businesses — I’m looking at $40,000-$50,000 per year in recovered time. The tools cost me maybe $40/month. The return on investment isn’t close.

But here’s what matters more: that 15-20 hours is time I get to choose how to spend. I spend it on strategy. On relationships. On building things. On thinking. That’s wealth I can’t measure on a spreadsheet.

The Philosophical Shift: From Tool to Team Member

Here’s the thing that took me longest to understand: the biggest impact of AI isn’t efficiency. It’s expanded capacity.

I used to think about this wrong. “How can I use AI to do more of what I’m already doing faster?” is the wrong question. The right question is: “What becomes possible if I have access to an extra 20 intelligent hours per week?”

The answer is: a lot.

For a solo founder or a small team, that’s not just an efficiency gain. That’s the difference between feasible and infeasible. Between possible and impossible. You go from being a one-person company constrained by one person’s time to being a company with leverage.

That shift has changed how I think about hiring, about which projects to pursue, about what scale is realistic. If I had to operate the way I did in 2022 — without AI — I would need at least two more people on my team. Not because the work is hard. Because there’s too much of it.

The other part of this shift is learning to think differently. The first few months, I was using AI like a tool: request, execute, move on. Now I’m using it like a thinking partner. I load it with context, I ask it questions that expose my own blind spots, I let it push back on my thinking, I iterate together.

That’s a different thing entirely.

When you have a smart advisor available instantly and constantly, your thinking gets sharper. You ask better questions because you have time to explore the space. You make fewer dumb mistakes because you have an external check on your ideas. You see around more corners because you’re thinking with someone who sees patterns differently than you do.

The professional world is going to bifurcate. People and teams that have integrated AI into their operating system will have impossible-to-match leverage. People and teams that are still operating 2022-style will find themselves increasingly constrained. I’m not being hyperbolic. The gap is that big.

Why Small Operators Win Most From AI

Here’s what’s counterintuitive: AI doesn’t help big enterprises as much as it helps solopreneurs.

Giant organizations with massive teams can optimize a 2% productivity gain across thousands of people and see real absolute returns. But they also have layers of management, process, and specialization that limit what AI can do. You can’t ask an AI to run a 500-person organization more efficiently. You can ask it to help write an email, but the upside is capped.

A solo operator or a small team? The upside is uncapped.

If I’m a one-person company and AI gives me the equivalent of two more team members, I just doubled my throughput. I can now take on projects I couldn’t before. I can execute faster than my competitors. I can maintain higher quality because I have cycles to think and iterate.

That’s the moment. That’s the advantage.

For me, running multiple businesses with relatively small teams, the AI advantage is that I can operate at much greater scale and complexity than I could hire for. I have the depth of thinking of someone with 15 years of experience across multiple domains. I have access to code-writing, design-thinking, copy-crafting, and strategy-building without hiring specialists for each.

If you’re already a large, well-resourced organization, AI optimizes what you’re doing. If you’re a small operator or a founder, AI expands what’s possible.

That’s why the next decade belongs to the people and small teams that embrace this, not the big established companies that treat it as a nice optimization.

What’s Next: Where This Is Heading

We’re still in year one of genuinely useful AI for serious work. And it’s already transformed how I operate.

What comes next is wild to think about.

The AI models keep getting smarter. Claude 3.5 Sonnet — what I use for 80% of my work — is already shockingly capable. Claude Opus is in a different league. In the next 12 months, the gap between what’s possible with AI and what’s possible without it will widen further.

The tooling is getting integrated. Claude Code makes the entire development loop faster. I can see a future where the entire workflow of building something is mediated by AI — where you think in concepts and the AI translates that to execution across multiple domains (code, design, copy, strategy).

The one thing I think people get wrong: you don’t need to be a prompt engineer. You don’t need to speak AI fluently or understand how it works under the hood. You just need to learn how to think with it. How to ask good questions. How to iterate. The best prompt writers I know aren’t prompt enthusiasts — they’re just clear thinkers who ask good questions. The AI fills in from there.

My advice to anyone watching this: start now. Not because you’re late — because you’re not. We’re still in the exponential growth phase where integrating AI into your workflow compounds month after month. In three years, everyone will be using it, and the edge you build now will be baked into how you think.

Find one workflow that genuinely bothers you — email drafting, research, code review, strategy thinking, whatever. Use Claude for a week. Get comfortable with it. Then move to the next thing. In three months, you’ll be operating differently. In a year, you won’t remember how you did things before.

The Broader Picture

What I’ve come to understand over these two years is that this isn’t hype. It’s infrastructure.

AI isn’t a tool that makes your existing workflow 10% more efficient. It’s a fundamental shift in how work happens. For small operators, for people trying to punch above their weight, for anyone who wants to compete with much larger, better-resourced teams, it’s become table stakes.

I run multiple revenue-generating businesses. I build side projects. I write consistently. I stay deeply involved in my team’s thinking. None of that is possible at my current scale without AI. Not because the work is hard, but because the work is volume.

The people I know who are most skeptical of AI are the ones with massive teams or the ones who haven’t integrated it yet. Once you use it seriously for three months, the skepticism evaporates. You realize you’ve been competing with a hand tied behind your back.

Here’s what I think is true: in five years, not using AI for business will be like not using email for business. It’s not a competitive advantage to use it — it’s a competitive disadvantage to not use it.

We’re in the window right now where early integration matters. Where you can build habits and workflows that will compound for years. Where the gap between AI-native operations and traditional operations is still visibly huge.

I’m taking full advantage of that window. If you’re running a business, you should too.