AI Self-Coaching · Prompts

Prompt Coaching:
50+ AI Coaching Prompts for Founders

Prompt coaching is only as good as the prompt. Here are 50+ specific prompts for pricing, churn, and acquisition — plus how to prompt an AI coach so it actually helps.

On this page
  1. What prompt coaching is (and isn't)
  2. Before and after: weak prompt vs coaching prompt
  3. Common mistakes when prompting an AI coach
  4. Pricing prompts
  5. Positioning prompts
  6. Acquisition prompts
  7. Churn prompts
  8. Strategy prompts
  9. Three founder outcomes
  10. Why good prompts aren't the whole answer

Prompt coaching is the practice of using a carefully written prompt to get real coaching out of an AI — not a generic answer, but a diagnosis and a recommendation you can act on. Done well, it turns a chatbot into a thinking partner. Done badly, it gives you a listicle you already could have written.

This page gives you 50+ prompt coaching templates for the questions solo founders actually face: pricing, positioning, acquisition, churn, and strategy. First, the part most prompt libraries skip — how to prompt so the answer is worth reading.

What Prompt Coaching Is (And Isn't)

Prompt coaching isn't asking "how do I grow my SaaS?" and hoping. It's giving the model your real numbers, your stage, and your constraint, then asking for one specific recommendation. The prompt does the work of turning a general model into a coach.

The difference is context and demand. A weak prompt asks for information. A coaching prompt hands over the situation and demands a decision.

Marcus
"I recommend judging any coaching prompt by one test: could a stranger answer it without knowing your business? If yes, it's too vague to coach you. Load it with your numbers and force a single recommendation."

Before and After: Weak Prompt vs Coaching Prompt

The same topic, prompted two ways. Watch what changes.

Pricing — before

Weak prompt

"How should I price my SaaS?"

This gets you a textbook: cost-plus, value-based, competitor-based, a paragraph on each. True, generic, useless for your Tuesday.

Pricing — after

Coaching prompt

"I charge $19/mo for a tool that saves freelance designers ~3 hours a week. I have 40 signups, 6 paying. I think I'm underpriced but I'm scared to change it. Give me one recommendation on price and the exact way to test it without losing the 6."

This gets you a diagnosis (underpricing signalled by the value gap), one recommended number, and a test plan. That's coaching.

Churn — before

Weak prompt

"How do I reduce churn?"

You already know the answers this returns: better onboarding, engage users, add value. No priority, no diagnosis.

Churn — after

Coaching prompt

"My churn is 9%/month. Most cancellations happen in the first 21 days. My onboarding is a single welcome email. Tell me whether this is an onboarding, value, or pricing problem — pick one — and give me the first thing to change."

The timing detail ("first 21 days") is what lets the model diagnose. Give it the pattern and it can name the problem type instead of listing all of them.

Common Mistakes When Prompting an AI Coach

Most disappointing AI coaching sessions fail for the same handful of reasons. Fix these and the quality jumps immediately.

  1. No numbers. "My growth is slow" can't be coached. "MRR flat at $1,400 for three months" can. Bring the real figures every time.
  2. Asking for options, not a decision. If you ask "what are my options," you'll get a menu and still have to choose. Ask for one recommendation and make the model defend it.
  3. Skipping your stage. Advice for a pre-launch founder and a $20k-MRR founder is opposite. State where you are or you'll get advice aimed at the average, which fits no one.
  4. Changing the question every message. Coaching compounds. Stay on one bottleneck until you have an action, instead of jumping topics each turn.
  5. Not telling it what you already tried. Without this, you get suggestions you've already exhausted. List what failed so it can skip past it.
  6. Accepting the first answer. Push back. "That assumes my traffic is the problem — it isn't, my conversion is" gets you a sharper second pass.

There's one mistake even a perfect prompt can't fix: a general model has no memory of your business between chats. You re-explain everything every time. That's the gap between prompt coaching and a purpose-built AI business coach that keeps your context. More on prompting Claude as a business coach if that's your tool of choice.

Pricing prompts

Pricing is where most solo founders leave the most money on the table. Use these prompts to pressure-test what you charge and how you package it.

Prompt

"I charge $X/month for [product]. My customers are [type]. Walk me through a value-based pricing check: what is the tool worth to them per month, and where does my price sit against that number?"

Prompt

"I have one flat price. Design a three-tier structure for [product] where the middle tier is the obvious choice. Tell me what to gate behind each tier and why."

Prompt

"I'm afraid to raise my price from $19 to $39. Give me the exact steps to test a price increase on new customers without touching existing ones."

Prompt

"Should I offer annual billing yet? Here are my numbers: [MRR, customer count, trial-to-paid rate]. Give me one recommendation, not a list of considerations."

Pricing prompts go further when the model understands value-based pricing first. Our how to price a SaaS product guide gives you the framework to bring into every one of these.

Positioning prompts

Positioning is the sentence that makes the right customer say "that's for me." These prompts force clarity.

Prompt

"Fill in this positioning sentence for my product and give me three versions: 'For [customer] who [problem], [product] is the only [category] that [differentiator].'"

Prompt

"My homepage headline is '[headline]'. Diagnose why it's vague and rewrite it so a stranger knows in five seconds who it's for and what it does."

Prompt

"I'm trying to serve everyone. Here's my customer list. Which single segment should I position around first, and what would I say to them specifically?"

Acquisition prompts

One channel that repeats beats five that don't. Use these to find and sharpen yours.

Prompt

"Here's where my last 10 customers came from: [list]. Identify the one channel worth doubling down on and tell me the next three actions to take on it."

Prompt

"I have no repeatable acquisition channel. Given my product [X] and my ideal customer [Y], recommend one owned channel to build first and why that one over the others."

Prompt

"Write me a 20-minute demo script that ends with a direct ask for payment, for [product] sold to [customer]."

For the full system behind these, see our guide on building a SaaS acquisition system.

Churn prompts

Churn quietly caps your growth. These prompts help you find the leak before you pour more into the top of the funnel.

Prompt

"My monthly churn is X%. Here's when customers tend to cancel: [timing]. Diagnose whether this is an onboarding problem, a value problem, or a pricing problem, and pick the most likely one."

Prompt

"Write me three cancellation-flow questions that will actually tell me why people leave, not just that they left."

Prompt

"My trial-to-paid rate is 4%. Walk me through the single most likely reason and one change to test this week."

When churn is the bottleneck, our guide on how to reduce SaaS churn pairs well with these prompts.

Strategy prompts

These are the zoom-out prompts for when you're stuck on what to work on at all.

Prompt

"Here's my situation: [MRR, stage, biggest frustration]. What is the single bottleneck I should fix first? Give me one answer and defend it."

Prompt

"I have 90 days. My goal is [goal]. Work backward and give me the one metric to move and the first action to take on it."

Prompt

"I keep adding features. Talk me out of it. What should I do instead this month given [numbers]?"

Three Founder Outcomes

Short, specific examples of prompt coaching applied to a real bottleneck. Numbers are illustrative of the pattern, not testimonials.

1. The underpriced tool

A founder charging $15/mo for a niche analytics tool used the pricing prompt above, brought real value numbers, and was pushed toward a $39 tier tested on new signups only. The point wasn't the price — it was that a specific prompt surfaced the underpricing they'd been avoiding for months.

2. The flat MRR plateau

A founder stuck at ~$2k MRR used a strategy prompt to force a single diagnosis instead of a to-do list. The prompt named acquisition — not product — as the bottleneck, which redirected weeks of feature work into one channel. See how to grow past an MRR plateau for the fuller playbook.

3. The early-cancellation churn

A founder with 9% monthly churn used the churn prompt with the timing detail included. The AI diagnosed a first-value problem, not a pricing one, and recommended a single onboarding change before touching anything else. One prompt reframed the whole plan.

Why Good Prompts Aren't the Whole Answer

Everything above works. A sharp prompt turns a general model into a genuinely useful thinking partner, and you should use these. But it's honest to say where the approach stops — because the prompt only ever fixes the input. It doesn't fix the four things that actually decide whether coaching changes your business.

1. Memory. A great prompt is only as good as the context you paste in that day. Next session, the model has forgotten your product, your numbers, and last week's decision — so you rebuild the setup every time, and the advice never compounds. The quality of the prompt can't solve the fact that there's no continuity behind it.

2. Accountability. A prompt gives you an answer and then goes quiet. It never asks whether you actually shipped the thing it recommended. Founders don't stall for lack of good advice — they stall because nothing follows up. No prompt, however well written, closes that loop.

3. Objectivity. You write the prompt, so you frame the problem — and you'll frame it around the answer you're already hoping for. Ask "how do I get more traffic" and you'll get traffic tactics, even when conversion was the real issue. A prompt obeys your framing. A coach is supposed to challenge it.

4. The discipline of one recommendation. A general model will happily hand you ten options because that feels helpful. Ten options is not a decision — it's homework. The hard part of coaching is committing to the single next move and defending it, and that's a design choice a prompt can't impose on the tool.

Marcus
"I recommend using these prompts — they're good, and better inputs beat worse ones every time. Just don't mistake a better input for a coach. The prompt is the question. It was never going to be the memory, the follow-up, or the nerve to give you one answer."

This is the honest line between prompt coaching and a coaching product. Prompts make any AI better at answering. They can't give it a persistent memory of your business, a reason to hold you to what you said, the distance to push back on your framing, or the discipline to stop at one recommendation. That's not a prompting problem you can write your way out of — it's what a purpose-built coach is for. GhostCoach runs the memory, the framework, and the single-recommendation rule underneath every session, so the input isn't the only thing carrying the weight. If you'd rather stay in your own tool for now, our guide to an AI business coach for solopreneurs shows what that structure looks like in practice.

Great prompts, zero memory. Marcus fixes the second part.

Every session already knows your product, stage, and 90-day goal — so you skip the setup and get one specific recommendation.

Talk to Marcus free → 14-day free trial · Builder from $79/mo · cancel anytime