How I Use AI to Do the Work of a 5-Person Team

I run four different businesses and consult on the side. My actual team? Three people. Five years ago, this setup would’ve required at least eight people to function at this level. AI isn’t the reason I pay less—it’s the reason I don’t need as much headcount in the first place.

This isn’t about replacing people or automating away jobs. It’s about fundamentally changing the economics of small teams and solo founders. Sam Altman was right when he said AI will enable a new generation of founders to operate at scales that used to require venture capital. I’m living that.

Key Points

AI replaces junior roles, not senior ones. Research analysts, junior copywriters, executive assistants, and entry-level developers are the positions that disappear—not strategic thinkers or creative leaders.
The time savings are real but not magical. AI cuts certain tasks in half, but you still need good judgment to review, edit, and integrate the output.
Quality is good enough for most work. 80% of tasks don’t require perfection. AI handles those 80% while you focus on the 20% that actually defines your brand and strategy.

The Five Roles I Don’t Hire For

When I started Rotate back in 2017, every agency had the same structure. You had partners at the top, senior strategists and designers in the middle, and a crew of junior people doing research, writing, scheduling, coding basics, and data entry. It was the only way to move fast. You needed bodies to do the repetitive work so senior people could focus on the strategic work.

AI changed that math completely. Let me walk through each role I no longer need to hire for, and what I do instead.

1. Research Analyst

What they did: Competitive analysis, market research, synthesizing data into digestable reports, tracking industry trends.

What I do now: I use Claude and ChatGPT to run competitive audits. I’ll throw a company name at it, ask for a breakdown of their positioning, pricing strategy, audience, and marketing angles. Then I take that 30-second output, spend 20 minutes on their website and Twitter, and have a solid competitive report that used to take a junior person three hours.

For deeper market research, I use AI to synthesize what I find. I’ll grab articles, market reports, and data points, dump them into Claude, and ask for key insights, trends, and gaps. The AI doesn’t replace the research itself—I still do that—but it saves the tedious work of reading through 50 sources and manually pulling out patterns.

Time saved per project: 8 hours to 2 hours. Quality? About 85% there, but honestly, for most competitive analysis, 85% is good enough if you spend an hour refining it.

2. Junior Copywriter

What they did: First drafts of email sequences, website copy, social media posts, ad copy, newsletter content.

What I do now: I use AI for all initial drafts. I’ll give Claude a brief—tone, audience, objective, key message—and it’ll write five email subject lines for our newsletter, or an entire cold email sequence, or a rewrite of a landing page headline. I spend 15 minutes giving feedback and refining, and it’s done.

Ryan’s Roundup, my newsletter with 12,500+ subscribers, would be impossible to write solo without AI. I sketch the key points, let Claude draft sections, I rewrite for voice, and we’re shipping. Before AI, I’d either hire a writer or spend all week on it myself.

The trickiest part is voice. Generic AI copy is useless—you have to train it on your actual writing, your personality, your quirks. Once I did that with Claude, the quality jumped from 60% to 85%. I’m editing, not rewriting.

Time saved per email/post: 45 minutes to 15 minutes. Quality: 80%, but the edits make it 100%.

3. Executive Assistant

What they did: Meeting prep, scheduling optimization, email triage, calendar management, expense reports.

What I do now: I’m not going to pretend I’ve eliminated this completely—I do some of it myself, and it’s still tedious. But I use AI to cut the time in half. Before calls, I ask Claude to summarize relevant emails, pull key context, and generate talking points based on previous conversations. For email triage, I use AI filters to sort critical messages from everything else.

Scheduling is where AI really shines. Instead of going back and forth with someone for 20 emails to find a time, I use a scheduling tool with AI-powered smart slots. It suggests times based on my calendar, timezone, and preferences, and most people book immediately.

The honest truth: this is the one role where I do notice the gap. An actual EA would catch things I miss and remember details I forget. But for a bootstrapped operation, the ROI on hiring a full-time EA doesn’t make sense when AI can handle 70% of the grunt work.

Time saved per week: 6-8 hours. Quality: 75%, but I’m okay with that trade-off.

4. Junior Developer

What they did: Prototypes, bug fixes, boilerplate code, basic integrations, tedious frontend tasks.

What I do now: I use Claude to write prototypes in a fraction of the time. I’ll describe what I need, and it’ll write React components, API integrations, or full-stack scaffolds. I review it, test it, and if there are gaps, I iterate with it in realtime. What used to be a 2-week sprint with a junior dev is now a 3-day sprint for me with Claude.

The speed is insane. Claude doesn’t get bored writing forms or CRUD operations. It doesn’t need breaks. It generates code that’s 80-85% production-ready. The remaining 15% is refactoring, testing, and edge cases that I actually care about fixing.

I’m not hiring junior developers anymore. If I need serious engineering work, I hire senior contractors who think, not code transcribers. Everyone in between is handled by AI.

Time saved per project: 2 weeks to 3-4 days. Quality: 80%, enough to ship.

5. Business Analyst

What they did: Financial modeling, KPI tracking, dashboard creation, report generation, data synthesis.

What I do now: I build financial models and dashboards with AI faster than I used to brief people on them. I’ll upload a spreadsheet, describe what I want to analyze, and Claude spits out formulas, what-if scenarios, and trend analysis. For dashboards, I use AI to write the queries and structure the logic.

Monthly reporting used to be a whole-day project. Now I dump the data, ask Claude to highlight what changed month-over-month, what’s trending up or down, and what should concern me. 20 minutes later, I’m reviewing the output and writing notes for the team.

The only real limitation here is that AI still makes mistakes with complex financial logic. You need to spot-check, especially if money is involved. But for operational metrics, KPI tracking, and insight generation, AI is honestly better than a human analyst because it’s faster and doesn’t get tired.

Time saved per report: 6 hours to 45 minutes. Quality: 85%, with verification required.

What AI Doesn’t Replace

Here’s the part where most AI hype falls apart: AI doesn’t replace senior people.

You still need someone who knows strategy, who can make judgment calls, who understands your brand and your customers at a deep level. AI is terrible at creative direction, at knowing which of five good options is the right option for your specific situation, at building trust with clients or partners.

In my experience, AI replaces the junior person doing repetitive work under supervision. It doesn’t replace the senior person doing the supervising. If anything, the leverage you get from senior people goes up when they have AI doing the grunt work. One strategist + AI beats one strategist + one junior every single time.

The other thing AI doesn’t replace is taste and intuition. My copywriter used to come with an instinct about what headlines work, what tone resonates, what positioning lands. AI has zero instinct. It’s statistical. I still need to apply judgment to every output.

And honestly? Relationships. AI can draft an email, but I’m still the one who has to send it, interpret the response, and maintain the relationship. That’s all me.

The Honest Economics

Let’s get specific about money. Here’s what this actually looks like:

The old way (hypothetical): Hire 8 people: 2 senior strategists, 1 account manager, 2 junior developers, 1 designer, 1 copywriter, 1 ops person. Plus benefits, taxes, recruiting costs. That’s easily $600K-$800K per year in payroll for a small operation like mine.

My way now: 3 employees on salary ($200K total), plus Claude Pro ($20/month), ChatGPT Pro ($20/month), Cursor ($20/month), a few other tools ($200/month). Total: ~$240K per year.

I’m not replacing eight people. I’m replacing four or five junior roles. The three people I have are strong—strategists, operators, people who can think independently. They’re doing the work that actually matters.

The Uncomfortable Question: What About Everyone Else?

I’m not naive about the implications. If more agencies and businesses operate like me, junior roles shrink. That’s a real shift in the job market, and it’s worth thinking about honestly.

My take: it’s not all bad, but it’s not all good either. The downside is obvious—fewer entry-level positions means harder breaks for junior talent. The upside is that the remaining positions demand better skills and pay better because they’re leverage-heavy.

I think the solution isn’t to resist this shift or pretend it won’t happen. It’s to accept it and get ahead of it. Juniors should skip the repetitive work and go straight to understanding strategy, judgment, and client relationships. Schools should teach people how to work with AI, not compete with it. And companies should be honest that the pyramid is flattening—fewer middlemen, more leverage for people who can think.

Personally, I’d rather work with three great people and a set of powerful tools than manage eight mediocre people. I think that’s the future. Three great people who are paid well, have clear leverage, and actually move the needle. That’s a better outcome than a bloated team where half the people are doing work that AI could do anyway.

How I Actually Use This

If you’re running a small business or a solo operation, here’s the practical playbook:

Map your team roles. Write down what your junior people actually do. Is it repetitive? Is it following a framework? Then AI can probably do 70-80% of it.

Start with the painful work. The first thing to automate is the task that everyone hates doing but everyone has to do—whether it’s report generation, email drafting, or competitive research.

Spend the time you save on judgment, not laziness. The time AI gives you back should go toward thinking strategically, deepening client relationships, or building product. Don’t just work less. Work differently.

Hire for senior roles. Stop hiring for the middle. Hire people who can think, critique AI output, and own outcomes. Pay them well. They’re your real leverage.

Have the voice conversation. If you’re using AI to write for your brand, train it relentlessly on your voice. Generic AI output is useless. Specific AI output that sounds like you is gold.

The Real Shift

The way I see it, AI isn’t eliminating the need for people—it’s eliminating the need for people doing repetitive work that doesn’t require judgment. The job market will shrink in some areas and grow in others. But for founders and small teams? This is our moment.

I can do work that used to require eight people because I have three exceptional people, good tools, and a willingness to spend 30 minutes a day reviewing and refining AI output. That’s the opportunity. Not “do everything yourself,” but “amplify the people you have.”

That’s how I’m running multiple businesses with a tiny team. Not because I’m superhuman. Because I’m using the right tools, focused on leverage, and ruthlessly honest about what requires human judgment and what doesn’t.