how I use AI to run a 7-figure business in 20 hours a week
most founders I know work 60-80 hours a week and still feel behind.
I used to be one of them. when Reeve and I scaled Funnel Duo Media past 8 figures, I was answering emails at midnight, reviewing ad copy at 5am, and spending weekends on reporting decks that nobody read twice.
today I work about 20 hours a week. revenue is higher. output is 3-4x what it was. and I actually see my daughter before she goes to bed.
this is not a flex. this is a structural change in how small businesses can operate, and most founders haven't caught up yet.
the shift that made this possible
in 2024, 40% of small businesses were using AI tools regularly. by 2025, that number hit 58%. solo-founded startups surged from 23.7% of all new companies in 2019 to 36.3% by mid-2025.
those numbers tell you something. the barrier to running a real business just dropped by an order of magnitude. the things that used to require headcount now require a system and a few hundred dollars a month in software.
but here is what the numbers don't tell you: most people using AI are using it wrong. they're adding tools on top of broken workflows. more output, same confusion.
what actually compresses your work week isn't AI. it's knowing which work to hand off.
the three categories
every task in my business falls into one of three buckets.
| Category | Examples | AI Handles? |
|---|---|---|
| Content | Research, first drafts, repurposing, scheduling, social media management | Yes, 80-90% |
| Operations | Email triage, reporting, client onboarding docs, invoicing, data entry | Yes, 70-80% |
| Analysis | Market research, competitor tracking, ad performance breakdowns, trend detection | Yes, 60-70% |
that leaves strategy, relationships, and creative direction. the 20% that moves the needle. the part I actually want to spend my time on.
before AI, I was spending 70% of my week on category 1 and 2. execution and admin. now those run in the background while I focus on the work that builds the business long term.
what the actual stack looks like
I'm not going to list 47 tools. most founders need fewer tools, not more.
here is what I actually use daily:
for content: AI agents handle research, first drafts, and platform-specific formatting. I review, add my perspective, and approve. what used to take me 4-5 hours per piece now takes 30-45 minutes of my time. the agents do the heavy lifting. I add the judgment.
for operations: automated email sorting and response drafting. client onboarding flows that trigger without me touching anything. reporting dashboards that update themselves and flag anomalies instead of making me dig through data.
for analysis: competitor monitoring that surfaces what changed this week instead of making me check 15 tabs. market research briefs that arrive ready to read, not ready to start.
the total cost of this stack runs between $500-1,000 a month. a single junior hire in Malaysia costs more than that. a comparable team in the US would run $15,000-20,000 monthly minimum.
what humans still need to do
this is the part most AI content leaves out.
AI is fast and cheap and consistent. it is not strategic. it does not understand why a client is nervous about a rebrand. it cannot feel that a market is shifting before the data confirms it. it does not build trust over a lunch meeting.
the 20 hours I work each week are spent on:
- strategy calls with clients. the work that retains accounts and deepens relationships. no AI can read the room the way a human can.
- creative direction. I set the angle, the positioning, the emotional tone. AI executes the brief. if the brief is wrong, the execution is irrelevant.
- relationship building. partnerships, collaborations, mentorship. the things that compound over years. you can't automate a handshake.
- quality control. the final eye. catching the thing that feels slightly off. the sentence that sounds like a robot wrote it. the strategy that technically makes sense but misses the human element.
- thinking. actual uninterrupted time to think about where the business is going. this is the one I undervalued for years. it might be the most valuable use of my time now.
why 20 hours, not zero
some people hear "AI automation" and think the goal is to do nothing.
that is how you end up with a business that produces content nobody reads, sends emails nobody opens, and makes decisions nobody thought through.
the goal isn't zero hours. the goal is to spend every hour on the work that actually requires you. the rest is infrastructure. let the machines run the infrastructure.
I set up a Guinness World Record event in December 2025 for the Largest AI Marketing Lesson. over a thousand people showed up. the entire event logistics, content preparation, and follow-up were AI-assisted. but the lesson itself, the message, the story, the human connection in that room. that was mine.
AI gave me the bandwidth to do something ambitious. it didn't do the ambitious thing for me.
your first moves
if you're a founder working 50+ hours and feeling stuck, here is where I'd start:
| Step | Time | Result |
|---|---|---|
| Track your time for one week. Write down every task. Be honest. | 5 min/day | Clarity |
| Sort tasks into the three categories (content, operations, analysis). Mark which ones only you can do. | 30 min | Framework |
| Pick one category. Automate the easiest task in it. Don't try to automate everything at once. | 2-3 hours | First win |
| Measure the time saved after 2 weeks. Reinvest that time into strategy or relationships. Not more execution. | 30 min | Proof |
| Expand to the next category. One task at a time. Build the system around your judgment, not around speed. | Ongoing | Compound |
the window is now
a complete solopreneur AI stack in 2026 costs between $3,000-12,000 per year. that is a 95-98% cost reduction compared to building a traditional team.
the founders who restructure their time around this reality in 2026 will have a compounding advantage that late movers cannot replicate. not because the tools won't still be available later. but because the strategic clarity and the systems thinking you build now compound every single month.
every week you spend doing work a machine could do is a week you didn't spend on the work only you can do.
I learned that the hard way. spent years grinding when I should have been thinking.
now I think for 20 hours a week and the business runs better than it ever did.
- Jackson