The Prompt Library I Use Every Day
After two years of daily AI use running a digital marketing agency in San Diego with 25-plus active clients, I have refined my prompts down to a core library. I use Claude primarily because the output quality is strongest. Most prompts also work well with ChatGPT.
Prompt 1: The Complete Ad Copy Generator
Structure: You are an expert Google Ads copywriter. I am advertising SERVICE to AUDIENCE in LOCATION. Differentiators: LIST. Write 25 headlines at 30 characters max and 10 descriptions at 90 characters max. Cover these angles: benefit-focused, problem-agitated, social proof, urgency, direct offer. Label each angle used. Why it works: Character limits and angle requirements prevent generic output.
Prompt 2: The Landing Page Copy Builder
Structure: I am building a landing page for OFFER targeting AUDIENCE. Conversion goal: ACTION. Objections to address: LIST. Proof points: TESTIMONIALS and STATS. Write a headline and subheadline, a 3-benefit section, an objection-handling section, social proof framing, and a CTA section. Why it works: Structuring the brief around objections forces persuasive rather than descriptive copy.
Prompt 3: The Competitive Analysis Brief
Structure: Here is the website copy for three competitors in INDUSTRY: PASTE COPY. Analyze what positioning each uses, what they do not address, and what gaps a competitor could own. Format as a structured brief with recommendations. Why it works: Comparative analysis with explicit output format produces actionable insights.
Prompt 4: The Email Sequence Creator
Structure: Write a 5-email nurture sequence for AUDIENCE who completed TRIGGER EVENT. Eventual offer: OFFER. Pain points: LIST. Goal: build trust and move toward CONVERSION ACTION. Emails 150 to 250 words each. Conversational tone. Include subject lines. Why it works: Specifying word count and tone prevents the AI from writing essays when you want nurture content.
Prompt 5: The Monthly Client Report Narrative
Structure: I am writing a monthly marketing report for CLIENT TYPE in INDUSTRY. Performance data: PASTE DATA. Write a 400-word executive summary explaining what happened in plain English, highlighting three wins, identifying the top challenge and our response, and setting expectations for next month. Why it works: The plain English requirement prevents jargon that makes reports unreadable for clients.
Prompt 6: The Blog Post Brief Generator
Structure: I need to write a blog post targeting KEYWORD. Audience: DESCRIPTION. I have RELEVANT EXPERIENCE. Write a detailed brief with a compelling headline, H2 and H3 outline, differentiating angle note, key points per section, and data to include. Why it works: Building the brief before writing improves structure and differentiation.
Prompt 7: The Google Business Profile Post Writer
Structure: Write 5 GBP posts for BUSINESS TYPE in LOCATION. Topics: LIST. Each 100 to 150 words, ending with a call to action, including an engagement question where natural, reflecting a specific brand voice. Avoid filler phrases. Why it works: GBP posts are a repeatable format and this prompt is optimized for it.
Prompt 8: The Objection-Handling Preparation
Structure: I have a sales call with PROSPECT TYPE. Likely objections: LIST. For each write the underlying concern, a bridge acknowledging their perspective, and a response addressing that concern. Under 60 words each. Tone: confident but not combative. Why it works: Framing objections as underlying concerns is the difference between defensive and consultative sales.
Prompt 9: The Service Page SEO Optimizer
Structure: Here is the current copy for a SERVICE page: PASTE COPY. Target keyword: KEYWORD. Audience: DESCRIPTION. Rewrite to naturally include the primary keyword, improve benefit clarity, add a section addressing the primary objection before hiring, and strengthen the CTA. Keep word count within 10 percent of original. Why it works: Constraining word count prevents unnecessary expansion.
Prompt 10: The Content Repurposing Factory
Structure: Here is a blog post I wrote: PASTE POST. Repurpose into 3 LinkedIn posts with different angles, 5 tweet-length posts, a 200-word email newsletter section, and a 90-second video script. Maintain my voice. Each format should feel native to its platform. Why it works: The native format requirement produces content that actually fits each format.
Getting the Most From These Prompts
These are starting points. Add more specific context about your business, audience, voice, and constraints and the output improves dramatically. For more AI workflow strategies, browse our blog or visit our services page. Ready to build a real AI-powered marketing workflow? Get in touch.
See also: about Derick Downs for background on how these prompts evolved from two decades of agency work, and how AI is changing SEO in 2026 for the broader context.
Frequently Asked Questions
What is the best AI tool for writing marketing prompts?
Claude is my primary tool for marketing prompts because it follows structured instructions more precisely and handles nuanced business contexts better than most alternatives. ChatGPT with GPT-4 is a strong second choice and often works well for quick ad copy iteration. The tool matters less than the quality of your brief — a specific, well-structured prompt produces far better output from either platform than a vague one does.
How do I write an effective AI prompt for Google Ads?
Specify character limits (30 for headlines, 90 for descriptions), list your differentiators and target audience, name the angles you want covered (urgency, social proof, direct offer), and tell the AI exactly how many variants to produce. Generic prompts produce generic ads. The more specific business context you provide, the more differentiated and usable the output becomes.
Can AI prompts replace a professional copywriter?
For first drafts of standard marketing assets — ads, emails, meta descriptions — AI prompts with a strong brief can match or exceed junior copywriter output at a fraction of the cost. For brand voice development, campaign concepting, and content requiring genuine subject matter expertise, experienced human writers still deliver better results. Use AI for volume and iteration; invest humans in strategy and differentiation.
How do I get AI to write in my brand voice?
Train the AI on your existing content by including examples in your prompt. Describe your voice explicitly — formal or conversational, direct or nuanced, industry jargon level. List phrases you use and phrases to avoid. The more voice examples and constraints you provide in the system prompt or project context, the more consistently on-brand the output becomes across every task.
What prompts work best for SEO content?
The blog post brief generator prompt works best for SEO content. Include your target keyword, your specific audience, your relevant experience or case studies, and a note about the differentiating angle you want to take. Ask for a detailed outline with data points per section before writing the full post. This produces structured content that covers topics comprehensively rather than surface-level.
How many AI-generated ad variants should I create for testing?
For a new Google Ads campaign, target 20 headlines and 6 descriptions as a starting point for a Responsive Search Ad. This gives Google’s optimization enough variants to identify winners. After 30 days, pull the asset report and cut low-performing variants, then use AI to generate replacements based on the patterns of what is working. Building a library of 60-plus headlines over time gives you real performance data to inform future creative direction.
