The AI Prompts I Actually Use to Run My Startup Alone

I run Super Prompts by myself. No co-founder, no marketing team, no copywriter on retainer.

Every week I ship features, write landing page copy, plan content, analyze what's working, and figure out what to build next. AI handles about 40% of that workload. Not because it's brilliant, but because the right prompt at the right moment means I skip the blank-page problem entirely.

These are the prompts I actually use. Not hypothetical. Not "50 prompts for everything." These are grouped by the moments where I'd otherwise get stuck, waste time, or pay someone I can't afford.

When You Need to Explain What You Built

The hardest part of launching solo isn't building. It's explaining.

You know what your product does. You know why it matters. But translating that into a sentence a stranger understands on first read? That takes longer than most features.

Positioning in one sentence

You are a positioning strategist. My product is [product] for [audience]. My competitors are [competitors]. Write a positioning statement that a customer would repeat to a friend. One sentence. No jargon. Focus on what's different, not what's better.

I run this prompt every time I update the landing page. The first output is never perfect, but it gives me three directions to react to instead of staring at a cursor.

Finding the gap your competitors miss

Compare [my product] to [competitor 1] and [competitor 2]. For each, list: what they do better than me, what they do worse, and the specific gap none of us are filling. Be honest about my weaknesses. Format as a table.

The honest part matters. If you don't tell the AI to be critical, you get cheerleading. I used this to find the angle for Super Prompts: nobody was building a visual prompt manager designed for creative workflows. Every competitor was a text dump.

Hero headline variations

Write 10 hero headlines for [product]. Under 8 words each. No jargon. Speak to [audience]'s biggest frustration. Rank by directness.

I usually throw away 8 of the 10. But the two that survive become my A/B test candidates. The constraint of 8 words forces clarity in a way that freewriting doesn't.

When You're the Entire Marketing Department

Solo means you're writing the tweets, planning the calendar, designing the emails, and deciding which channel to even bother with. These prompts compress that work.

Choosing where to spend your 5 hours

I'm launching [product] for [audience]. I have $0 marketing budget and 5 hours per week. Rank these channels by expected ROI for my first 100 users: Twitter/X, Reddit, Product Hunt, SEO content, cold email, communities, YouTube. Explain each ranking in one sentence.

This prompt saved me from spreading thin across six channels. The output consistently says the same thing: go where your audience already hangs out and contribute before you promote. Not revolutionary, but hearing it ranked by effort/return ratio helps.

A month of content in 15 minutes

Create a 4-week content calendar for [product] targeting [audience] on [platform]. 5 posts per week. Alternate between: pain point, tip, social proof, behind-the-scenes, and direct CTA. Write the hook for each post.

I don't use the hooks word-for-word. But having 20 starting points means I spend my creative energy refining instead of generating. That's the real leverage here.

The landing page you'd pay $2K for

You're a conversion copywriter. Structure the messaging for [product]'s landing page: hero headline, subheadline, 3 benefit blocks, social proof section, objection handling, and final CTA. Write the actual copy for each. My audience is [audience] and their biggest frustration is [frustration].

The specificity of "biggest frustration" changes the entire output. Without it, you get generic SaaS copy. With it, you get something that sounds like it was written for your actual users.

When You Need Words That Sell

Writing copy is a different skill than writing code or writing emails. These prompts bridge that gap.

Turning features into feelings

Here are the features of [product]: [list features]. For each, write the benefit in this format: "Feature → So you can → Which means". The final "which means" should be emotional, not functional.

This is one of my most-used prompts. "Folder organization" becomes "so you can find your best prompts in seconds, which means you stop recreating work you've already done." That emotional hook is what landing pages need and feature lists miss.

Welcome emails that don't feel automated

Write a 3-email welcome sequence for new signups of [product]. Email 1: quick win in 60 seconds. Email 2: deeper feature on day 2. Email 3: social proof plus upgrade nudge on day 5. Under 150 words each. Conversational, not corporate.

The "quick win in 60 seconds" instruction is everything. Most onboarding emails explain. The good ones activate.

Handling objections before they become churn

List the top 5 objections someone would have before signing up for [product]. For each, write a 1-2 sentence response for a landing page FAQ. Be honest, not dismissive.

I add "be honest, not dismissive" because AI defaults to spin. When a user asks "Why should I trust a new tool?" the answer shouldn't be "We're the best!" It should acknowledge that trust is earned and explain how.

When You Need to Decide What to Build

This is where prompts stop being about content and start being about product decisions. These save me from building the wrong thing.

Ruthless prioritization

I'm a solo founder of [product]. Here's my feature backlog: [list features]. My current users are: [count] with [usage pattern]. Prioritize using ICE scoring (Impact, Confidence, Ease). Format as a ranked table with scores.

I run this quarterly. The AI doesn't know my business better than I do. But it forces me to think in terms of impact per hour, which is the only metric that matters when you're one person.

Cutting scope before you overcommit

I'm planning to build [feature]. Here's the full spec: [describe feature]. As a product advisor, tell me: what's the absolute minimum version that would validate the hypothesis? What can I cut without losing the learning? Be ruthless.

"Be ruthless" is a magic instruction. Without it, AI hedges. With it, you get the version of your feature you should've scoped from the start.

Knowing your one metric

I'm tracking growth for [product]. Define the one metric that matters most at my current stage: [user count] users, [revenue] MRR, [growth rate]. Explain why this metric, how to measure it, what "good" looks like, and what to ignore right now.

The "what to ignore" part is the most valuable output. Early on, everything feels important. This prompt cuts through that noise.

When You Need to Create Content That Attracts Users

Content marketing as a solo founder means writing something useful, not just writing something.

Blog post outlines that don't read like AI

Write a detailed outline for a blog post titled "[title]" targeting [keyword]. Include: H2 headings, key points under each, one unique angle competitors aren't covering, and a CTA that connects to [product]. Target: 1,500 words.

The "unique angle" instruction matters. Without it, you get the same outline everyone else has. With it, you at least get a direction to push against.

Twitter threads that feel earned

Turn this idea into a Twitter/X thread: [idea]. Hook tweet with a curiosity gap. 5-7 value tweets, one idea each. Closing tweet with CTA. Under 280 chars each. Make it feel like earned insight, not content marketing.

"Earned insight, not content marketing" is the difference between a thread that gets saved and one that gets scrolled past.

Re-engagement without guilt

Write an email to users who signed up for [product] but haven't logged in for 7 days. Acknowledge they're busy. Remind them of the one thing they were trying to accomplish. Give them one link. Under 100 words. No guilt.

"No guilt" changes the whole email. AI defaults to "We miss you!" which nobody wants to read. This produces something human.

When You Need to Find Users

Growth prompts for when you have zero budget and need to be creative.

Cold outreach that isn't spam

Write a cold DM to [target person type] about [product]. No fake personalization. Lead with a specific observation about their work. Explain what I built in 1 sentence. Ask a genuine question. Under 80 words.

I've sent maybe 30 messages based on variations of this prompt. Response rate: around 25%. The "specific observation" instruction is what makes it work. Generic outreach gets ignored.

Community launches that don't get flagged

I'm launching [product] in [community/subreddit]. Write a launch post that: shares a genuine story about why I built this, provides value without the product, mentions the product naturally, and asks for feedback. Follow [community]'s posting culture.

"Follow the community's posting culture" is crucial. Reddit hates self-promotion. Indie Hackers rewards transparency. The AI adjusts tone when you tell it the norms.

SEO content briefs

Write a content brief for a blog post targeting "[keyword]" for [product]. Include: search intent, target word count, required H2 sections, questions to answer, competing content to beat, and the unique angle.

I use this before writing every blog post, including this one. It's faster than manual SERP analysis and gives me a structure to react to.

Making These Prompts Compound

The real shift happens when prompts build on each other.

Run the positioning prompt first. Feed that positioning into the headline prompt. Use the winning headline in your content calendar prompt. That chain produces better results than running any prompt in isolation, because each step carries forward the context from the last.

The problem is losing track of which version worked. I've burned hours recreating a prompt I used three weeks ago because it was buried in a chat log.

That's why I built Super Prompts. It's a visual prompt manager where you save the prompts that actually work, organize them by project or workflow, and pull them up the next time you need them. Instead of recreating your best work, you iterate on it.

It's free to start. The prompts above are a good beginning. Saving the ones that work for your specific product is what turns them from a list into a system.