5 AI Prompts That Actually Move the Needle for Founders

Most founder productivity prompts are too generic to be useful. “Help me prioritize my tasks” produces a list of tips anyone could have written. These five prompts are designed differently — they are specific enough to generate output you can actually act on today.

Each one targets a real bottleneck: a task that is frequently delayed, frequently done badly, or rarely attempted at all.

Prompt 1: The Weekly Reality Check

Use this on Monday morning before your week starts.

I am a [stage] founder building [product]. Here is what I said my top priority was
last week: [priority]. Here is what I actually spent most of my time on: [list].

Be direct with me: was last week's time allocation aligned with that priority?
If not, name the most likely cause — be specific, not generic.
Then give me one structural change to this week's plan (not a mindset change)
that would close the gap. Do not give me a list. Give me one thing.

Why it works: The constraint “one thing” prevents the output from becoming a multi-point advice list you will skim and forget. The instruction to be specific rules out the most common AI response pattern — validating your effort while offering vague suggestions.

Prompt 2: The Pre-Mortem for Any Major Decision

Use this before committing to a significant choice — a hire, a pivot, a pricing change.

I am about to [decision]. My reasoning for doing this is: [reasoning].
My biggest concern is: [concern].

Steelman the case against this decision as strongly as possible.
Then give me the three questions I should be able to answer before I commit,
and identify the single assumption my reasoning depends on most heavily.

Why it works: Founders are prone to confirmation bias on their own decisions. This prompt explicitly inverts the reasoning — not “help me think this through” but “argue against me.” The best-case output is an objection you had not considered.

Prompt 3: The Sell Domain Activation Prompt

Use this when you have been spending too much time in Build and your outbound has gone quiet.

I run [product] targeting [customer segment]. My current offer is [offer].
I have not done outbound in [N] days. Here are the last two customers I closed
and why they said yes: [reasons].

Write me five personalized outreach opening lines targeting [specific company type]
that reference something specific about the recipient's situation — not generic value props.
Assume the recipient is a [job title] at a company that [specific situation].

Why it works: It forces you to provide the context that makes personalization possible, which is the step most founders skip. The output is five opening lines you can evaluate and adapt, not a template you have to make specific yourself.

Prompt 4: The Spec That Gets Built Right

Use this when you are handing a feature to a contractor or describing something to an engineer.

I need to build [feature]. The problem it solves for the user is: [problem].
The user is [user type] and their current workflow is: [workflow].

Write a one-page spec that includes:
1. Problem statement in user terms
2. What "done" looks like (acceptance criteria)
3. The three most likely edge cases
4. What this feature explicitly does NOT do (scope boundary)
5. The one thing that would make this spec fail in implementation

Keep each section to three sentences or fewer.

Why it works: The “what this explicitly does not do” section is what prevents scope creep. The “what would make this fail” section surfaces the assumption the spec is hiding, which is usually where implementation goes wrong.

Prompt 5: The Quarterly Goal Translation

Use this at the start of each quarter to connect a big goal to this week’s work.

My most important goal for this quarter is: [goal].
This week I have [N] hours of focused time available across Build, Sell, and Operate.

Break this quarter's goal into the three milestones that, if hit, guarantee the goal.
Then tell me what I need to do this week — specifically and concretely — to make
progress on the nearest milestone. Prioritize ruthlessly. I should end with no more
than two things on my list.

Why it works: The constraint “no more than two things” is the most important part. Most planning prompts produce exhaustive lists. This one produces a ranked short list, which is the output that actually guides daily work.

Your action for today: Copy Prompt 1 into Claude right now and fill in the blanks for your last week. That is it.

Related:

• The Complete Guide to AI Tools for Founders
• What AI Actually Delivers for Founders
• Planning with Claude AI

Tags: AI prompts for founders, Claude prompts, founder productivity, weekly planning prompts, AI decision-making