ChatGPT Prompts for Marketing That Actually Work

Why Most ChatGPT Marketing Prompts Fail

I've watched thousands of agency owners and founders try to shortcut their marketing with AI. Most of them end up frustrated - not because ChatGPT is bad, but because they're prompting it like a search engine instead of a strategy partner. They type in "write me a cold email" and get something that sounds like it was written by a robot who's never made a sales call in its life.

The quality of what comes out is entirely determined by the quality of what goes in. Vague prompt, vague output. If you give ChatGPT a role, a specific prospect, a concrete pain point, and a word limit - the output is actually usable. That's the difference between a tool that saves you an hour a day and one that wastes it.

The numbers bear this out. Research consistently shows that prompt quality increases output quality by five to fifteen times over weak prompting. And with three out of four marketers now using ChatGPT at least weekly for work tasks, the gap between people who know how to prompt and people who don't is showing up directly in results - in open rates, in booked calls, in content that ranks versus content that sits idle.

This guide breaks down the exact prompt frameworks I use across every major marketing function: cold email, ICP research, content, SEO, social media, video scripts, objection handling, and lead research. These aren't copy-paste generic prompts from a listicle. These are the structures that produce first drafts you can actually send, publish, or deploy.

The Anatomy of a High-Quality Marketing Prompt

Before diving into specific prompts, you need to understand what makes a prompt work. Every strong ChatGPT marketing prompt has five components:

• Role: Tell ChatGPT who it is. "You are an experienced B2B sales copywriter who specializes in cold outreach for SaaS companies." This anchors the voice and point of view.
• Context: Give it your product, your target company size, their industry, and the specific pain point you're solving. No background info means generic output.
• Constraint: Set a word count, tone, and format. "Under 120 words, conversational, no corporate jargon, plain text output."
• Goal: Be explicit. "The goal is to start a conversation - not pitch the product. End with a one-line question."
• Output format: Specify exactly what you want back. "Return a subject line and email body in plain text."

That five-part structure is what separates a usable output from something you'll delete and rewrite from scratch anyway. Master the structure, then customize the content for each use case below.

One more thing before we get into the actual prompts: ChatGPT is a drafting engine, not a replacement for your judgment. It can produce volume and structure at speed. It cannot replace your firsthand knowledge of your market, your product, or your customer. Use these prompts to handle the structural heavy lifting - then add the human layer that makes copy actually land.

ChatGPT Prompts for Cold Email Marketing

Cold email is where I've spent most of my career - writing it, testing it, and teaching it to 14,000+ agencies and founders. ChatGPT can draft a solid first version of a cold email in under two minutes. But raw AI output always needs a human layer on top. It can't personalize at scale, and prospects can spot generic AI copy immediately. Your job is to use it for the structural heavy lifting, then add one real, specific detail about the recipient before you hit send.

Cold Email First Draft Prompt

Use this prompt to generate a first draft for any outbound campaign:

"You are a B2B sales rep at [your company]. Write a short cold email to a [job title] at a [company size]-person [industry] company. They are struggling with [specific pain point]. We help companies like theirs achieve [specific outcome]. Keep the tone conversational and human - no buzzwords or sales fluff. Under 110 words. Output a subject line and email body in plain text. End with a single low-pressure question."

The more specific the pain point, the better the output. "Struggling with low reply rates on outbound campaigns" beats "struggling with sales" every time. If you want a library of tested cold email frameworks to pair with this, grab my free Cold Email GPT Prompts pack - it has the prompt structures I've refined across hundreds of real campaigns.

Follow-Up Email Prompt

Most deals happen in the follow-up, not the first touch. Use this:

"Write a short follow-up email to someone who didn't respond to my first cold email. Reference that I reached out previously without being passive-aggressive. Add one new piece of value - either a case study, relevant stat, or industry insight for [their industry]. Keep it under 80 words. Friendly and direct. End with a different CTA than the original email."

Re-Engagement Email Prompt

For contacts who went cold after showing initial interest:

"Write a re-engagement email to a prospect who had a discovery call with us [timeframe] ago but went silent. Don't blame them for not responding. Reference the problem we discussed and offer one new angle they might not have considered. Under 90 words. Conversational. End with a simple yes/no question."

The key with re-engagement is giving them a new reason to respond rather than just checking in. ChatGPT is good at generating new angles when you give it the original pain point - feed it context from your notes on the original conversation and it'll surface approaches you wouldn't have thought of on your own.

If you're running these sequences at scale, you'll want to pair the AI drafts with a proper sending tool. Smartlead and Instantly both let you rotate email variants automatically across a sequence - run your ChatGPT-generated options as A/B tests and you'll have real data on what resonates within days, not weeks.

ChatGPT Prompts for ICP and Market Research

Before you write a single email or ad, you need to know exactly who you're talking to. ChatGPT is excellent for rapid ICP (Ideal Customer Profile) development - especially when you're entering a new vertical or testing a new offer.

ICP Definition Prompt

"Act as an experienced B2B marketing strategist. I sell [your product/service] to [broad target market]. Help me define my ideal customer profile. Include: company size, industry, likely job titles I should target, their top 3 business pain points, what they care about most when evaluating a solution like mine, and what objections they typically raise. Be specific - avoid generic answers."

This gives you the foundation for every downstream marketing asset: your cold email, your LinkedIn content, your ad copy, your landing page. Don't skip this step. Want to go deeper on market research with AI? My free GPT Market Research Prompts pack covers competitor analysis, positioning, and buyer psychology prompts in detail.

Pain Point Mining Prompt

"You are a [job title] at a [company type]. What are your top 5 frustrations with [category your product solves]? What language would you use to describe these problems to a colleague? What would it take for you to book a 15-minute call with a vendor claiming to solve this?"

Read the output and mine it for exact language to use in your copy. The phrases ChatGPT generates when it's roleplaying your buyer are often closer to actual prospect language than anything you'd come up with internally.

Competitor Analysis Prompt

Understanding how your competitors position themselves is essential for finding your angle. Use this:

"Act as a B2B market analyst. I compete with [list competitors] in the [industry] space. Based on how these companies typically position themselves, identify the messaging gaps - what pain points are they NOT addressing that a buyer in this space would care about? What unique angle could I own that they're leaving on the table? Be specific and give me 3-5 distinct positioning opportunities."

Feed the output into your cold email hooks, your landing page headline, and your LinkedIn bio. The goal is to find the claim none of your competitors are making confidently - that's where you plant your flag.

ChatGPT Prompts for SEO and Content Marketing

Content marketing for B2B is about getting found by buyers before they're actively shopping - and then staying top of mind until they are. ChatGPT can dramatically accelerate the content creation process if you know how to prompt it correctly. Marketers using ChatGPT for blog posts, social media, and email report saving an average of 52% in production time - that's a real efficiency gain you can redirect into strategy.

Keyword Research and Clustering Prompt

ChatGPT can't pull live search volume data - you still need Ahrefs or Google Search Console for that. But once you have a list of keywords, ChatGPT is excellent at organizing them into a logical content strategy:

"Here is a list of 30 keywords: [paste list]. Group them into topic clusters based on semantic relevance and shared search intent. For each cluster, suggest: a pillar page title, 3 to 5 supporting article ideas, and the primary keyword for each. Output as a structured list."

This approach - clustering related keywords under a pillar structure - signals depth to search engines and improves your chances of ranking across multiple terms simultaneously. It also prevents keyword cannibalization, which quietly tanks a lot of agency and founder blogs.

Blog Outline Prompt

"Act as a content strategist for a B2B [industry] company. Create a detailed blog post outline targeting the keyword '[your keyword]'. The post should address search intent for someone at the [awareness/consideration/decision] stage. Include: H2 sections, suggested subheadings, key points to cover in each section, and one data point or example per section. Target 1,500 words."

One thing I do that most people don't: I specify the search intent stage explicitly. A post written for someone who's never heard of cold email needs a completely different structure than a post written for someone comparing cold email tools. Getting that wrong means you're writing content that doesn't match what the searcher actually needs - and that kills your ranking potential regardless of how good the writing is.

Meta Title and Description Prompt

On-page SEO is where most people leave ranking potential on the table. Use this to generate variations fast:

"Write 5 SEO-optimized meta titles and meta descriptions for a page targeting the keyword '[primary keyword]'. Keep titles under 60 characters and meta descriptions under 155 characters. Focus on clarity, relevance, and click-through rate. The audience is [describe your target reader]. Avoid clickbait - make each one accurate to the page content."

Pick the strongest two or three options and test them against each other. Even a 5% improvement in click-through rate from the SERP compounds significantly over time - it's one of the highest-leverage on-page optimizations you can make without touching a single backlink.

Content Refresh Prompt

Got a post that ranked well but has started sliding? This is often more effective than writing new content:

"I have a blog post targeting '[keyword]'. Here is the current content: [paste]. Identify sections that are outdated, too thin, or missing intent alignment. Suggest specific additions, rewrites, and structural changes to make this post more comprehensive and search-engine competitive. Keep the existing strong sections intact."

Refreshing an existing page that Google already knows about is almost always faster than building authority on a brand new URL. Use this prompt quarterly on your top-traffic pages.

FAQ Section Prompt

FAQ sections help you pick up featured snippets and People Also Ask placements without writing a whole new article:

"Act as an experienced SEO copywriter. Generate a comprehensive FAQ section for a blog post titled '[title]'. Include 8-10 questions that a reader at the consideration stage would actually ask. Write the answers in plain language - no jargon. Keep each answer between 40 and 80 words. Structure as H3 questions with paragraph answers."

ChatGPT Prompts for Social Media Marketing

Social media for B2B is a long game. The goal isn't viral reach - it's consistent visibility in front of the specific buyers who will eventually convert. ChatGPT can accelerate your content calendar significantly once you give it the right constraints.

LinkedIn Post Prompt

"Write a LinkedIn post for a [your industry] founder. The post should open with a one-sentence hook that makes someone stop scrolling. The topic is [specific insight or lesson]. Write in a direct, conversational tone - no fluff, no corporate language. No bullet-point lists. Under 200 words. End with a question that invites comments."

Add 2-3 examples of your own previous top-performing posts and tell ChatGPT to match that style. The more style guidance you give it, the less it defaults to generic LinkedIn-speak. If you want to accelerate your LinkedIn scheduling and analytics beyond manual posting, Taplio integrates AI-assisted writing with scheduling and engagement tracking - a solid pairing with these prompts.

LinkedIn Carousel Prompt

Carousels consistently outperform plain text posts in organic reach on LinkedIn. Use this structure:

"Create a 7-slide LinkedIn carousel on the topic '[topic]'. Each slide should have a headline (under 10 words) and 2-3 supporting bullet points. Slide 1 should be the hook - make it a bold, counterintuitive statement or a compelling question. The final slide should have a clear CTA. Audience: [describe your ICP]. Tone: direct and practitioner-level, not academic."

Twitter/X Thread Prompt

"Write a 10-tweet thread on '[topic]'. Tweet 1 is the hook - it should make someone want to read the rest of the thread. Each subsequent tweet should deliver a standalone insight. Keep each tweet under 250 characters. No filler tweets that just say 'let's get into it.' End with a CTA tweet that links to [your resource]. Tone: direct, confident, no corporate buzzwords."

Content Calendar Prompt

Use this to batch a full month of content in one session:

"Act as a social media strategist for a B2B [industry] company. Create a 4-week LinkedIn content calendar. Include: post topic, format (text, carousel, video script, poll), the core insight or angle for each post, and the buyer intent stage it targets (awareness, consideration, decision). The audience is [describe ICP]. Mix educational content, case study angles, and direct-value posts. Output as a table."

Batch your content calendar prompts once a month, generate the drafts in one sitting, and schedule them out. That's how you stay consistent without content creation consuming your entire week.

ChatGPT Prompts for Video Marketing Scripts

Video is one of the highest-leverage content formats for B2B - especially on YouTube and LinkedIn. If you're not creating video content, you're leaving a massive visibility channel untapped. ChatGPT can draft a solid script structure in minutes, which removes the biggest friction point most people have: the blank page.

YouTube Video Script Prompt

"Write a YouTube video script on the topic '[topic]'. Target length: [X] minutes (approximately [X x 130] words). The audience is [describe viewer]. Structure: open with a hook that states the problem the viewer has, then deliver [number] actionable steps or insights, then close with a CTA to [subscribe/download resource/book a call]. Tone: direct and conversational, like talking to a colleague. No intro music placeholders or generic transitions."

The best YouTube scripts I've written start with the viewer's problem, not with who I am or what the video is about. ChatGPT defaults to the latter unless you tell it otherwise. Override that default every time.

Short-Form Video Script Prompt (Reels, Shorts, TikTok)

"Write a 60-second video script on '[topic]'. Hook in the first 3 seconds - use a bold statement, a surprising stat, or a direct question. Deliver one core idea with a real example. End with a single CTA. Conversational, zero fluff. Output as a script with [PAUSE], [VISUAL CUE], and [B-ROLL] markers where relevant."

Video Repurposing Prompt

If you already have video content, ChatGPT can help you extract more value from it:

"Here is a transcript of a [length] video: [paste transcript]. Help me repurpose it into: (1) a LinkedIn post, (2) a 3-tweet thread, (3) a blog post outline, and (4) 5 short-form video clip ideas with their hooks. Keep the original voice and insights intact."

This is the content multiplication play. One video becomes five or six assets across different channels, all rooted in the same core idea. Your time investment stays the same - your content footprint expands significantly.

ChatGPT Prompts for Lead Generation

ChatGPT can help you think through your prospecting strategy, but it can't actually pull real contact data for you - that's where dedicated tools come in. The combination of AI-generated targeting criteria and actual B2B data tools is where the real leverage lives.

Use this prompt to sharpen your targeting before you start pulling lists:

"I sell [product] to [industry]. Help me identify 5 specific signals that would indicate a company is a strong fit and likely has budget and urgency right now. Think about: hiring patterns, tech stack, recent news, funding events, or operational changes. For each signal, suggest how I could find companies exhibiting that signal."

Once you have your targeting criteria locked, you need actual contact data. ScraperCity's B2B lead database lets you filter prospects by job title, seniority, industry, location, and company size - so you're pulling a list that matches the exact ICP your ChatGPT prompt just helped you define. That pairing is the whole game: AI for targeting logic, a real database for execution.

If you need to find someone's direct contact details once you've identified a target, this email lookup tool will handle the search without manual digging. And if your campaign requires direct dials rather than email, ScraperCity's Mobile Finder surfaces direct phone numbers for the contacts on your list.

Prospect Research Prompt

Once you have a list of target accounts, use this prompt to quickly build context before outreach:

"Act as a sales researcher. I'm about to reach out to [company name], a [company type] that [brief description]. Research context to help me personalize outreach: What are the most likely pain points for a [job title] at this type of company? What recent company triggers (funding, hiring, product launches, leadership changes) would make them more receptive to [your offer]? What angle would resonate most with someone in this role?"

Use this one prospect at a time for high-value accounts, or adapt it to generate a general profile for your target persona that your whole team can reference. For more prompt frameworks specifically around lead gen and prospecting, grab my free GPT Lead Gen Prompts pack.

Technographic Prospecting Prompt

If your offer pairs well with specific tech stacks, this is a high-signal targeting approach:

"I sell [product] to companies using [specific tool or tech stack]. Generate 10 outreach angles that reference their tech stack as a context point - not in a creepy way, but as a signal that I understand their setup. Each angle should connect their tech choice to a pain point my product solves. Keep each angle to 2-3 sentences."

This kind of technographic targeting pairs naturally with a tool like ScraperCity's BuiltWith Scraper, which identifies the tech stack of any website so you can build lists of companies running the specific tools your product integrates with or competes against.

ChatGPT Prompts for Ad Copy

Paid advertising lives and dies by the copy. ChatGPT can generate enough ad variations in ten minutes to keep your creative team busy for weeks - and the cost-per-test drops to almost nothing when you're iterating on AI-drafted copy instead of paying a copywriter to write every variant from scratch.

Facebook and Instagram Ad Prompt

"Write three variations of a Facebook ad for [product/service] targeting [audience]. Each variation should test a different psychological trigger: one focused on fear of missing out, one on social proof, and one on a specific outcome the buyer wants. Keep each ad under 150 words. Include a headline, primary text, and CTA for each."

Google Search Ad Prompt

"Write 5 Google Search ad variations for the keyword '[target keyword]'. For each ad, write a headline 1 (under 30 chars), headline 2 (under 30 chars), headline 3 (under 30 chars), and a description line (under 90 chars). Focus on benefit-led headlines, not feature-led. The audience is [describe]. Product: [describe]. Primary CTA: [describe]."

Retargeting Ad Prompt

"Write 3 retargeting ad scripts for people who visited my [product/service] landing page but didn't convert. Each variation should address a different reason someone might have hesitated: (1) trust/credibility concern, (2) price/value uncertainty, (3) timing - not sure if now is the right moment. Under 120 words each. Include headline, body copy, and CTA."

Retargeting copy is fundamentally different from cold traffic copy - you're writing to someone who already knows what you do. Telling ChatGPT that context explicitly produces much better output than treating it like a first-touch cold audience.

ChatGPT Prompts for Email Newsletter Marketing

A newsletter is one of the highest-value assets you can build as an agency owner or founder. It's direct access to your audience without a platform algorithm in the way. ChatGPT can handle the heavy lifting on structure and drafting so you can focus on the insights only you can provide.

Newsletter Edition Prompt

"Write a B2B newsletter edition on the topic '[topic]'. The audience is [describe readers - their role, company type, and core interest]. Structure: open with a 1-2 sentence hook referencing something happening in the industry right now, then deliver 3 practical insights with examples, then close with one recommendation or resource. Under 500 words. First-person voice, no corporate tone. No filler phrases like 'in conclusion' or 'in today's fast-paced world.'"

Newsletter Subject Line Prompt

"Generate 10 subject lines for a B2B newsletter edition about [topic]. Mix approaches: curiosity-based, direct-value, question-led, and bold contrarian take. Under 50 characters each. No exclamation points. No clickbait. The audience is [describe]."

Your newsletter subject line is your open rate. Everything else is downstream. Run 3-4 of these through your email platform's A/B test feature before committing to one. If you're not already using an email platform with proper list segmentation and automation, AWeber handles the fundamentals cleanly for most B2B newsletters at the scale most founders are operating at.

ChatGPT Prompts for Objection Handling

One of the most underused applications of ChatGPT in marketing and sales is objection prep. Instead of getting blindsided on a call or in an email reply, you can use AI to war-game every objection your prospects might raise before it happens.

Objection Roleplay Prompt

"Act as a skeptical [job title] at a [company type] who is evaluating [your product]. I'm going to pitch you. After my pitch, respond with the 5 most common objections someone in your position would have. For each objection, explain the underlying concern behind it - not just the surface-level pushback."

Then follow up with:

"Now help me write a concise, non-defensive response to each of those objections. The tone should be consultative and confident - not salesy. Focus on addressing the underlying concern, not just answering the surface question."

This exercise alone is worth doing before any new campaign launch. Feed the output into your email sequences, your sales scripts, and your FAQ pages.

Sales Script Prompt

For discovery calls, use this to build a call framework rather than a rigid script:

"Act as an experienced B2B sales trainer. Help me build a discovery call framework for selling [product] to [job title] at [company type]. Include: 3 opening questions to establish rapport and diagnose the situation, 4 pain-discovery questions that uncover the real business impact of [problem], 2 questions to understand decision-making process and timeline, and a transition line to move from discovery into a solution overview. Format as a logical flow, not a rigid script."

Proposal and Pitch Deck Outline Prompt

"Create a structured outline for a sales proposal for [your service] targeting a [company type]. The prospect's main challenge is [pain point]. Include sections for: executive summary, understanding of their situation, proposed solution, deliverables, timeline, social proof/case study, investment, and next steps. For each section, provide 2-3 bullet points of content guidance."

If you want AI-generated proposal templates you can drop directly into a client-facing document, check out my free Proposal AI Templates resource - it covers the full proposal structure with fill-in-the-blank AI prompts for each section.

ChatGPT Prompts for Email Subject Lines

Subject lines are where most outbound campaigns live or die. A great email body is worthless if nobody opens it. Use ChatGPT to generate volume quickly, then test the winners.

"Generate 15 subject lines for a cold email targeting [job title] at [industry] companies. The email is about [your offer/topic]. Mix the approaches: some curiosity-based, some direct and benefit-led, some referencing a specific pain point. Keep all subject lines under 8 words. No clickbait. No exclamation points."

Then pick the 3-4 that feel most natural and A/B test them inside your sending tool. If you're using Smartlead or Instantly, both platforms let you rotate subject lines across a sequence automatically - pair that with your ChatGPT-generated variants and you'll have data on what resonates within days, not weeks.

Personalization Hook Prompt

Subject lines work even better when paired with a personalized first line. This is the one piece of every email you should never fully automate:

"Here is publicly available information about this prospect: [paste their LinkedIn about section, recent post, or company news]. Write 5 personalized opening lines for a cold email that reference this information naturally - not in a way that feels like stalking. Each line should be under 20 words and transition smoothly into a business conversation."

Use this for your top 20-30 target accounts where manual personalization is worth the time investment. For everyone else, the templated approach works - just make sure the pain point is specific enough to feel relevant even without personal detail.

ChatGPT Prompts for Campaign Planning and Strategy

Beyond individual copy assets, ChatGPT is a genuinely useful strategic partner for campaign planning - especially for thinking through multi-channel sequences and messaging frameworks you haven't tried before.

Campaign Brief Prompt

"Create a campaign brief for an outbound marketing push targeting [audience]. Here is the goal, offer, and context: [describe]. Include sections for: primary objective, target audience profile, core message and differentiator, channel mix (email, LinkedIn, ads, content), asset list needed, success metrics, and timeline. Keep each section tight - this is a working brief, not a deck."

Multi-Channel Sequence Prompt

"Design a 14-day outbound sequence for reaching [job title] at [company type]. The offer is [describe]. Map out every touchpoint across email, LinkedIn, and cold call. For each touchpoint specify: the channel, timing, message goal (not full copy), and the CTA. Assume no response until I note otherwise. Prioritize variety of channel and angle across touchpoints - no two consecutive touches should use the same approach."

This multi-channel framework is how serious outbound teams operate. Single-channel campaigns miss the buyers who don't check email but live on LinkedIn, and vice versa. If you want to automate the LinkedIn side of this, Expandi handles LinkedIn outreach automation while staying within safe usage limits - pair it with your ChatGPT-designed sequence for a coordinated multi-touch approach.

Messaging Framework Prompt

"Build a messaging framework for [your product/service]. Include: (1) primary value proposition in one sentence, (2) three supporting proof points with evidence, (3) the top three buyer personas and the specific angle that resonates with each, (4) the core objection for each persona and the counter-narrative, (5) three alternative ways to describe what we do for different contexts - sales call, cold email, LinkedIn post. Output as a structured document."

This is the document your whole team should have on day one. It means everyone is using the same language, the same angles, and the same objection responses - rather than each person winging it with their own interpretation of what the product does.

Advanced Prompting Techniques for Marketers

Once you've got the basic prompt structures down, there are a few advanced techniques that dramatically improve output quality without adding much time.

Chain Prompting

Instead of trying to get everything in one massive prompt, break complex outputs into a chain of connected prompts. Start with strategy, then ask for copy, then ask for variations. Each step gives the AI a clearer, narrower task - and the cumulative output is much stronger than a single all-in-one prompt would produce.

Example chain for a cold email campaign: (1) Define the ICP and pain points. (2) Generate 10 possible angles for outreach. (3) Pick the 3 strongest angles and write a first-touch email for each. (4) Write follow-up emails for each angle. (5) Generate 15 subject line variants across all three email versions.

Style Injection

If you want ChatGPT to match a specific voice or writing style, paste in 2-3 examples of content written in that style and ask it to analyze the characteristics before generating new content. This is much more effective than describing the style in abstract terms like "conversational" or "professional."

"Here are three examples of my writing style: [paste examples]. Analyze the tone, sentence length, vocabulary, and structural patterns. Then write [new piece of content] matching that style exactly."

Constraint Testing

When you're not sure what format will work best, generate multiple constrained versions and compare. Ask for one version under 80 words, one under 150, one as a list, one as a narrative. The version that feels most natural to you is usually the one that will resonate most with your audience.

Fact-Check Layer

Never publish AI output without a fact-check pass. ChatGPT can confidently state incorrect statistics, attribute quotes to the wrong people, or generate plausible-sounding but fabricated case studies. Build a review step into every AI-assisted workflow: human edits for accuracy, then a final pass for personality before anything goes out the door.

How to Build a Reusable ChatGPT Prompt Library

The biggest efficiency gain isn't writing one great prompt - it's building a library of proven prompts your whole team can pull from. Every time a prompt produces a strong output, save it in a shared doc with the context variables clearly marked in brackets. Anyone on your team can fill in the variables and get a consistent, usable result without starting from scratch.

Organize your library by function: cold email, follow-ups, LinkedIn content, ad copy, ICP research, objection handling, SEO, video scripts, newsletter. Add notes about which prompts work for which industries or offer types. Over time, this becomes one of your most valuable internal assets - a documented system for producing marketing output at speed.

Here's the structure I use for every saved prompt in my own library:

• Prompt name: Short label for quick reference (e.g., "Cold Email First Draft - SaaS")
• Use case: When to use this prompt and what output it produces
• The prompt: Full text with [variables] clearly marked
• Best for: Industries, personas, or offer types where this prompt works best
• Output notes: What to expect, what to edit manually, what not to trust the AI on

That simple structure is enough. You don't need a fancy tool for this - a shared Google Doc or Notion page works fine. The point is that the library exists and is accessible, not that it lives in a sophisticated system.

Team Prompt Governance

If you're running a team, someone needs to own the prompt library. Designate one person to review new prompts quarterly, retire ones that have stopped producing strong output, and update existing prompts based on what's working in the field. Treat it like any other operational asset - it degrades without maintenance.

The one thing ChatGPT still can't do is replace your judgment about what's actually true about your market, your product, and your customers. Always fact-check AI output before publishing or sending, and always add at least one human-specific detail before any outbound message goes out. The AI handles the volume; you handle the accuracy and the personality.

What ChatGPT Can't Do for Your Marketing

I want to be direct about the limits here, because I've seen too many people burn time and money treating ChatGPT like a magic solution rather than a drafting tool.

ChatGPT has no access to real-time data. It cannot tell you your domain authority, current keyword rankings, live ad performance, or backlink profile. You still need Ahrefs, Google Search Console, and your ad platform dashboards for that. It also has no understanding of your specific brand voice, your client history, or what's actually been working in your outbound campaigns - unless you tell it explicitly in the prompt.

It also can't replace the relationship layer in sales. It can draft the email, but it can't read the room on a discovery call. It can write the LinkedIn post, but it can't respond authentically to the comment thread. The human judgment layer is still the most valuable thing you bring to your marketing - AI just makes the execution layer faster and cheaper.

The teams seeing the strongest results are treating ChatGPT as a skilled but uninformed drafting partner: excellent at structure and volume, dependent on you for accuracy, voice, and relevance. Give it that context and it delivers. Ask it to work without context and you'll get the generic output that's giving AI-assisted marketing a bad name in a lot of circles.

If you want me to walk through how I use these prompts inside an actual outbound system - from ICP definition through to booked calls - that's what I cover inside Galadon Gold.

Quick-Reference: ChatGPT Marketing Prompt Templates by Use Case

Here's a condensed reference table of the core prompts in this guide. Copy these into your prompt library and fill in the brackets for your specific situation.

• Cold Email First Draft: Role as B2B sales rep + pain point + outcome + under 110 words + single question CTA
• Follow-Up Email: Reference first touch + one new value add + different CTA + under 80 words
• ICP Definition: B2B strategist role + product/market + company size + pain points + objections
• Pain Point Mining: Roleplay as buyer + top frustrations + language they use + booking threshold
• Competitor Analysis: Market analyst role + competitor list + messaging gaps + 3-5 positioning opportunities
• Keyword Clustering: Paste keyword list + cluster by intent + pillar page + 3-5 supporting articles
• Blog Outline: Content strategist role + keyword + intent stage + H2/H3 structure + data points per section
• Meta Titles and Descriptions: 5 variations + under 60 chars title + under 155 chars description + CTR focus
• LinkedIn Post: Industry founder voice + hook sentence + insight + under 200 words + engagement question
• LinkedIn Carousel: 7 slides + hook slide + 2-3 bullets per slide + CTA slide
• YouTube Script: Problem hook + actionable steps + CTA close + conversational tone
• Short-Form Video: 3-second hook + one idea + example + single CTA + visual cue markers
• Facebook Ad: 3 variations + FOMO / social proof / outcome triggers + headline + body + CTA
• Google Search Ad: 5 variations + 3 headlines under 30 chars + description under 90 chars
• Objection Roleplay: Skeptical buyer roleplay + 5 objections + underlying concerns + non-defensive responses
• Cold Email Subject Lines: 15 variants + mixed approaches + under 8 words + no exclamation points
• Campaign Brief: Objective + audience + message + channel mix + metrics + timeline
• Multi-Channel Sequence: 14-day map + email/LinkedIn/call touchpoints + channel + timing + CTA per touch
• Newsletter Edition: Industry hook + 3 insights + close recommendation + under 500 words
• Content Calendar: 4-week LinkedIn plan + format + angle + intent stage + output as table

Every one of these prompt structures is a starting point, not a finished product. The best version of any prompt you use will be the one you've iterated on three or four times based on the outputs you've gotten. Save the evolved version, not the original. That's how you build a library that actually compounds in value over time.