Prompts that sell: AI-powered copywriting that converts

The problem: AI generates text, not sales

You're using ChatGPT or Claude to write your sales pages, launch emails, and LinkedIn posts. The text is fine—smooth, error-free, readable. But the conversions aren't there.

The problem? You're asking AI to write, when you should be asking it to sell. And selling isn't just about stringing words together. It's about touching emotion, creating desire, handling objections, driving action.

The good news: with the right prompts, AI can become your most effective copywriter. Here's how.

Mistake #1: asking for "marketing content" without a framework

Weak prompt:

Write a sales page for my project management SaaS

This prompt gives you generic text with no punch that looks like everyone else's.

Prompt that sells:

You're a B2B SaaS copywriter.

Context:
- Product: Project management SaaS for technical teams
- Audience: CTOs and dev leads at 10-50 person startups
- Main problem: losing 10hrs/week in pointless status meetings
- Unique solution: auto-generated dashboards from Git commits

Write the hero section (headline + subheadline + CTA) using the PAS framework:
1. Problem (pain point experienced)
2. Agitate (amplify the impact)
3. Solve (present the solution)

Tone: direct, technical but accessible, no marketing BS.

The difference? You're providing a proven copywriting framework (PAS), specific context, and a brand voice.

The 3 copywriting frameworks to use in prompts

1. AIDA (Attention, Interest, Desire, Action)

Perfect for sales emails and landing pages.

Write a launch email for [product] following AIDA:

A - Attention: open with a shocking stat about [problem]
I - Interest: explain why current solutions fail
D - Desire: tell the story of a customer who solved this problem with [product]
A - Action: single CTA to [link] with urgency argument (limited offer)

Length: 150-200 words max
Tone: conversational, like a friend giving advice

2. Before-After-Bridge

Ideal for social posts and Facebook ads.

Create 3 variations of a LinkedIn post for [product/service] using the Before-After-Bridge structure:

Before: describe the current frustrating situation of [target audience]
After: paint the picture of the ideal situation
Bridge: present [product] as the path between the two

Each variation must:
- Start with a 10-word max hook
- Use bullet points for readability
- End with an engaging question
- Be 100-120 words

3. Storytelling + objection handling

Powerful for long-form sales pages.

Write a customer case study following this structure:

1. Initial situation: [customer] had [specific problem + numbers]
2. Failed attempts: list 2-3 solutions tried without success
3. Discovery: how they found [product]
4. Implementation: the first 3 concrete actions
5. Results: before/after numbers at 30, 60, 90 days
6. Customer quote: authentic statement about benefit #1

Naturally address these objections:
- "It's too expensive"
- "I don't have time to implement this"
- "I've already tried similar tools"

Format: 600-800 words, clear subheadings, authentic testimonial tone

Pro tip: create copywriter personas

Instead of re-explaining the tone every time, define reusable personas.

You are "Sarah", a B2B SaaS copywriter.

Writing style:
- Short sentences (15 words max)
- No corporate jargon (avoid: "innovative solution", "market leader")
- Use concrete numbers, not vague adjectives
- Always include a tangible benefit in every paragraph
- Direct tone like between professionals

Specialties:
- Transforming features into emotional benefits
- Handling objections with social proof
- Creating urgency without being manipulative

Now, as Sarah, write [specific request]...

Save this persona. Every time you need sales copy, simply recall "You are Sarah" and give your instruction.

The 5-question test before validating AI copy

Before publishing your AI-generated copy, check:

1. Does the headline grab attention in 3 seconds? If you have to think about it, rewrite.
2. Is the problem described using words your customer actually uses? Not jargon only you understand.
3. Is there at least one concrete number or social proof? "Save time" doesn't sell. "Save 8hrs/week" does.
4. Is the call-to-action clear and singular? One CTA per piece of copy.
5. Would your grandmother understand what you're talking about? If not, simplify.

If even one answer is no, rework your prompt.

Complete example: prompt for a sales email sequence

Create a 4-email sequence to convert newsletter subscribers into paying customers.

Product context:
- Online course "AI Copywriting" at $297
- Audience: freelancers and content creators already using ChatGPT
- Offer duration: 5 days
- Early bird bonus: 30 ready-to-use copywriting prompts (value $97)

Sequence structure:

Email 1 (Day 0, immediate send):
- Launch announcement
- Framework: Before-After-Bridge
- CTA: discover the program

Email 2 (Day 2):
- Case study: how [persona type] increased conversion rate by 34%
- Handle objection: "I already know how to use ChatGPT"
- CTA: join the course

Email 3 (Day 4, 24hrs before end):
- Urgency reminder + early bird bonus
- List 3 concrete benefits as bullet points
- Short testimonial from beta participant
- CTA: enroll before expiration

Email 4 (Day 5, last 3hrs):
- Short email (80 words max)
- Direct tone, no storytelling
- Just: time remaining + link + signature

Overall tone: direct, friendly, zero manipulative pressure.
Emails 1-3 length: 200-250 words.

For each email, also suggest 3 subject line variations (50 characters max).

How KayaPrompt simplifies this work

If you're thinking "OK, but I don't want to recreate these prompts every time"... you're right.

That's exactly why we built KayaPrompt: a library of ready-to-use copywriting prompts, organized by objective (sales email, landing page, social post, etc.).

You fill in the variables (your product, your audience, your offer), and the system generates the complete prompt with the right framework.

No need to reinvent the wheel for every campaign. You keep what works, iterate on the rest.

The real secret of AI copywriting

AI doesn't replace a good copywriter. It amplifies their method.

If you master the frameworks (AIDA, PAS, Before-After-Bridge), if you know your audience, if you can structure a sales argument... AI becomes a machine for producing variations to test.

But if you just ask it "write me something that sells," you'll get something that doesn't sell.

The difference between average AI copy and AI copy that converts? The quality of the prompt.

Your turn

Pick a sales text you wrote recently (email, post, landing page). Redo it with one of the frameworks above. Compare the results.

And if you want to save time on this part, explore the ready-to-use copywriting workflows at kayaprompt.com. We've done the structuring work for you.