Solo Founder Workflow

The Solo Founder's 20-Minute Claude Stack (Morning to Ship)

April 22, 2026 · 8 min read

Most "AI productivity" posts assume you have a team, a backlog, and someone else worrying about payroll. That's not you. You're the founder, the designer, the copywriter, the support desk, the person who fixes the Stripe webhook, and the person who remembers to buy coffee. You have maybe 90 focused minutes before the first "quick question" email fires.

This is the 20-minute Claude routine I run before I touch any real work. Three timed blocks. No fluff. No "chat with your AI coach." Just the part of the day where Claude earns its keep — and the part where it doesn't.

Fair warning: if you're expecting a magical prompt that turns Claude into your employee, this isn't that. Solo founders don't need an employee. They need a force multiplier that handles the 40% of the workday that isn't actually the work — the thinking, the structuring, the first drafts. That's what this is.

Why 20 minutes, and why in blocks

I've tried the "use Claude all day" approach. It fails for the same reason "write all day" fails: thinking has a ceiling and you hit it fast. After 25 minutes of continuous prompt-and-iterate, output quality drops off a cliff and you don't notice until you ship something mediocre.

What actually works: short, deliberate blocks with clear handoffs. Three of them, 6–7 minutes each, with real deliverables at the end of each one. The moment the block is done, Claude goes back in the drawer and I get on with the work.

Claude is not your morning coffee. It's a specific tool you pull out for specific moves. If it's always open in a tab, you're doing it wrong.

Block 1 — The Agenda Block (6 minutes)

Before I do anything, I need to know what today's work actually is. Not the to-do list — the one thing that, if shipped, makes today not wasted. Solo founders tend to treat their days as a list of obligations instead of a sequence of bets. That's what kills momentum.

I start here every morning with the same prompt. Yes, the same one. Don't get clever:

You are helping a solo founder set today's single shipping bet.

My context:
- Product: [one line — what I sell, to who]
- This week's goal: [one sentence]
- Yesterday I did: [3 bullets, honest, including the things I avoided]
- Meetings today: [list or "none"]
- Energy level, 1–10: [number]

Task:
1. Look at yesterday's list and flag anything I'm avoiding because it's scary vs. just annoying.
2. Propose ONE shipping bet for today — a concrete thing that leaves the laptop as a finished artifact (published post, sent email, deployed feature, scheduled demo, etc).
3. Propose 2 "next best" alternatives in case the first one is blocked.
4. Give me a time estimate for each, realistic not optimistic.

Write it like a no-nonsense COO. Don't coach me. Don't tell me I'm doing great.

The "don't coach me" line is not a joke. If you don't explicitly tell Claude to skip the encouragement, it will start every morning by telling you that you're making excellent progress. Great, now I feel vaguely infantilized and I still don't know what to ship.

What you want out of this block: one sentence that starts with "Today I ship…". If you don't have that sentence at minute 6, you didn't do the block right — you asked Claude for a plan and got a committee memo.

Block 2 — The Draft Block (8 minutes)

Now the actual shipping bet has a first draft. This is the block where Claude does the most work and you do the least. The trick is knowing exactly what to hand off and what to keep for yourself.

Hand off to Claude:

• First draft of the email / post / landing copy / spec
• A second draft in a different voice (so you can A/B against your gut)
• A list of 3 titles, 3 subject lines, or 3 CTAs
• A bulleted "what would a skeptic say about this" audit

Do not hand off:

• The final decision about tone or framing (that's your job)
• Anything involving your actual pricing or positioning without you reviewing every word
• "What should I build next" — Claude is not a product manager, and it doesn't know your customers

The prompt I reuse here is brutally plain. Solo founders overcomplicate prompts because they want to feel like they're "using AI properly." You're not. You're trying to ship.

You're drafting [email / landing page section / LinkedIn post / tweet thread].

Audience: [one sentence — who they are and what they want]
Goal: [what they should do after reading — click, reply, sign up, nothing]
Tone: [3 adjectives, e.g. "direct, warm, specific"]
Length: [word or line target]

Two constraints:
- No generic openings ("in today's fast-paced world", "let's dive in", etc.)
- No hedging ("can help you", "might be useful") — write it as if I know it's true.

Give me two versions. Version A: closer to my tone. Version B: 20% bolder.

The "Version A / Version B" split is the single best trick I've stolen for solo founder work. You always think you know your tone until you see 20% bolder sitting next to it. Nine times out of ten, the bolder one is the right call and you were self-editing your way into something forgettable.

Block 3 — The Close-Out Block (6 minutes)

Most solo founders skip this block and that's why half their work doesn't move the needle. You shipped the thing. Now you need the distribution layer, the small-print fix-ups, and the handoff to tomorrow.

I run three micro-tasks in this block, all with Claude, all in the same thread:

1. Distribution variants. "Take this post and give me (a) a LinkedIn version under 1,300 characters, (b) a tweet-thread version, 5–7 posts, (c) a 60-second reel script in monologue form." This is 2 minutes of work, saves you an hour later in the week, and is the single reason my content ratio looks like I have an agency.
2. The skeptic pass. "Read this like someone who hates marketing emails. What's the weakest line? What's the one sentence that makes them roll their eyes?" I edit based on the answer. Sometimes I delete the entire thing. It's fine.
3. Tomorrow's starter. "Based on what I did today, what's the obvious next shipping bet for tomorrow?" Save that answer in a note. Tomorrow's Block 1 starts at minute 2 instead of minute 0.

Six minutes. Every single day. This is the block that turns a solo founder who "works hard" into a solo founder who actually compounds output.

What Claude is terrible at (and don't keep trying)

Three things I've wasted hundreds of hours trying to force Claude into. If you're doing any of these, stop today:

1. Strategy for your specific business

Claude doesn't know your customers. It knows a general statistical shape of "SaaS customer" or "agency client." For anything that depends on your specific data — what churn looks like, which tier converts, which feature your power users actually use — you are the intelligence layer. Claude can structure your thinking (great). It can't replace your thinking (don't try).

2. Writing in your voice from zero examples

"Write this in my voice" produces generic corporate voice every time unless you give it 3–5 pieces of your actual writing first. If you don't have time to paste examples, don't use Claude for that piece. Write it yourself and use Claude on the next one. This is non-negotiable.

3. Deciding between two serious options

Pricing changes, hiring decisions, which feature to cut. If it's a real decision with trade-offs, Claude will always split the difference and produce a polite "both have merit" answer. Use it to list trade-offs. Don't use it to pick.

The full stack, on one screen

Here's what the 20 minutes look like, copy-paste-able:

06:00 — Coffee
06:05 — Block 1 (Agenda Block): 6 min
        → output: one "Today I ship…" sentence

06:12 — Block 2 (Draft Block): 8 min
        → output: Version A + Version B of the main artifact

06:21 — Block 3 (Close-Out Block): 6 min
        → output: distribution variants + skeptic pass + tomorrow's starter

06:28 — Claude off. Work on.

The rest of the morning is you, your actual editor, your actual code editor, your actual customers. That's by design.

Why this isn't "use AI more"

The pattern you're going to see from every AI tool this year is "stay in our interface longer." More chat, more agents, more integrations. For a solo founder, that's a trap. You are not trying to automate yourself into a zombie who watches Claude type.

You are trying to compress the 40% of your day that isn't the work — so you can spend the other 60% actually doing the work. The 20-minute stack is the smallest unit of AI I've found that survives a real solo founder week without turning into a distraction.

If you want the agents behind this workflow pre-built — the Monthly Content Calendar, the Landing Page Writer, the Weekly Business Review — that's exactly what the Business Builder Kit is. Seven agents, one time, ready to drop into your 20 minutes. No subscription, no seat fees, no upsell loop.