Case Study No. 2

AI Agent that transforms topics into Blogs, Captions, and Emails.

I’m looking for a marketing automation specialist who can help me batch and schedule monthly content: blog posts, emails, and social, in a way that’s strategic, seamless, and most importantly, aligned with my brand voice.

I've already gone through three contractors and wasted time and money. I need someone who has done this successfully before, can show examples, and doesn’t need constant handholding. There’s a partial system already in place, but you’ll assess whether it can be improved or should be rebuilt for better efficiency and results.

Scope of Work
Create and schedule 4 blog posts per month

The Solution

A fully autonomous content automation agent built using n8n, Google Sheets, OpenAI, and native integrations with Facebook, LinkedIn, X (Twitter), Gmail, and Google Sheets.

Starting from a single topic input, the agent automatically generates a blog post, tailored captions for each social platform, and a marketing email with subject and body content. The system runs end-to-end with minimal manual input, enabling consistent, high-quality content delivery across all channels.

Set-up

Before proceeding with solution building, curate the Google Sheet to serve as the central database for data storage.

Start by accessing Google Sheets and creating a new spreadsheet. Within the sheet, create a table with curated columns for data storage in each section. Required columns include:

1

Topic

Manually inputted (or automated) topics by the user.

2

Date Generated

Controls the scheduling of when the topic will be processed into content.

3

Blog Content

Stores the GPT-4o generated blog content.

4

Facebook Caption

Stores the GPT-4o generated Facebook caption.

5

Linkedin Caption

Stores the GPT-4o generated LinkedIn caption.

6

X Caption

Stores the GPT-4o generated X (formerly Twitter) caption.

7

Email Content

Stores the GPT-4o generated email content.

8

Email Subject

Stores the GPT-4o generated subject line for the email.

9

Status

A checkbox to indicate if content has been successfully generated.

10

Posted

A checkbox to monitor and control duplicate posting and sending.

Building the Solution

1Topic Retrieval from Google Sheets

n8n automatically retrieves blog topics marked as ready from a centralized Google Sheet, based on the scheduled date specified by the user. While initial topics are manually curated by the client, the system is built for flexibility, with future plans to integrate sources like Google Trends or RSS feeds for automated topic generation.

For each scheduled topic, the AI-powered content agent generates a complete set of deliverables: a blog post, platform-specific captions (Facebook, LinkedIn, X), an email subject line, and email body.

The workflow is configured to process up to four content sets per month, ensuring consistent output while allowing room for review and strategic adjustments.

In order to properly fetch the date for the topic to be generated, make sure to use and add the following code in an n8n Code node:

How Prompt Engineering powers this AI Agent

What makes this solution effective is not just the use of GPT 4o, but the intentional design behind each prompt.

Every step in the workflow uses structured instructions that are zero shot, role specific, and guided by clear constraints. These prompts are thoughtfully connected to lead the AI through a consistent and high quality content creation process.

As a result, each output aligns with the client's brand voice and meets professional expectations without the need for manual editing or supervision.
Interested? Let's talk

2Blog Generation with OpenAI GPT-4o

The selected topic is passed to GPT-4o using a structured prompt that frames the model as a professional blog writer. The AI generates a blog post between 400 and 800 words, written in a clear, strategic tone that aligns with the client’s brand voice. The word count is flexible and can be adjusted based on specific content requirements.

To ensure high-quality output without relying on static templates, the agent leverages prompt engineering techniques such as role instruction, tone guidance, and structured formatting. These techniques help maintain consistency, reinforce brand alignment, and meet content standards across all generated posts.

You are a professional blog writer.

Blog Generation: Smart Prompting

Thinking like a strategist.

Your prompt looks like:

You are a professional blog writer.

Write a short-form blog post between 400–800 words.

Structure the blog post like this: A compelling introduction, 2–3 core sections, and a short conclusion.

Use a tone that is natural, clear, and slightly conversational. Avoid bullet points and listicles.

Output only the blog text — no titles, explanations, formatting, or markdown.

Topic: {{ $json.Topic }}

Prompt Breakdown

Prompting Technique

How it's used?

Role Instructioning
The AI is assigned the role of a professional blog writer, shaping its behavior and output style to match expert level writing standards.
Content Constraint
The AI is instructed to write a blog post between 400 and 800 words, setting clear boundaries on the desired length.
Structural Prompting
The prompt specifies an ideal blog layout including an introduction, 2 to 3 core sections, and a short conclusion to guide the AI’s paragraph flow.
Tone Conditioning
The prompt defines the tone as natural, clear, and slightly conversational and restricts formats like bullet points and listicles to maintain consistency
Output Formatting
The AI is instructed to return only the plain blog text with no titles, explanations, formatting, or markdown.
Input Referencing
The prompt includes "Topic" using the expression 
{{ $json.Topic }} to feed the exact topic into the generation process ensuring relevance and contextual accuracy.

3Generate Platform-Specific Social Captions with OpenAI

To ensure that social posts match the client’s tone and feel natural on each platform, the agent uses tailored AI personas guided by custom style rules. Each platform prompt is crafted with a defined role, tone, and strict output constraints. This eliminates the need for manual rewriting or formatting, keeping the brand voice consistent across all channels.

Platform

AI Role

Style Rules

Output Constraints

Facebook
Social Media Copywriter
Friendly and conversational tone, include 2 to 3 emojis, end with a casual call to action
Return only the caption. No formatting or extra text.
Linkedin
Professional Content Strategist
Confident and insightful tone, open with a strong statement, end with a reflective or engaging question, include 2 to 3 hashtags
Return only the caption. No links, emojis, or formatting.
X (Twitter)
Digital Strategist
Short and punchy, highlight a single insight or teaser, keep under 280 characters, use up to 2 emojis and hashtags
Return only the caption. No links or formatting. Character limit enforced.

You are a Social Media Copywriter for Facebook

Facebook Caption

Smart Prompting

Your prompt looks like:

You are a social media copywriter for Facebook.

Your task is to write a single Facebook caption based on the blog post below.

Make it hook the reader, sound friendly, and end with a casual CTA.

Use 2-3 emojis. Add relevant hashtags.

Return only the caption.

Blog post: {{ $json.message.content }}

Prompt Breakdown

Prompting Technique

How it's used?

Role Instructioning
The AI is assigned the role of a social media copywriter for Facebook, aligning its response with platform-specific expectations and tone.
Task Specification
The AI is told exactly what to do—write a single caption based on a blog post—clearly defining scope and output.
Tone Conditioning
The prompt instructs the AI to make the caption hook the reader, sound friendly, and end with a casual call to action, ensuring a light and social tone.
Content Enhancement
The AI is asked to include 2 to 3 emojis and relevant hashtags, optimizing the content for engagement and platform appeal.
Output Formatting
The AI is instructed to return only the caption, removing any risk of unnecessary explanations, formatting, or extra content.
Input Referencing
The prompt uses "Blog post" with the expression 
{ $json.message.content }} as the input source, ensuring the caption is directly based on the original content.

You are a Professional Linkedin Content Strategist

Linkedin Caption

Smart Prompting

Your prompt looks like:

You are a professional LinkedIn content strategist.  

Your task is to write a single LinkedIn caption based on the blog post provided below.  

Begin with a thought-provoking hook, and maintain a confident, professional tone throughout.  

Do not include any links. Use relevant hashtags to enhance visibility.  

Return only the caption—no formatting, no markdown, and no additional commentary.  

Blog post: {{ $('Blog Creation').item.json.message.content }}

Prompt Breakdown

Prompting Technique

How it's used?

Role Instructioning
The AI is assigned the role of a professional LinkedIn content strategist, guiding its tone and output for a platform specific audience
Task Specification
The prompt clearly instructs the AI to write a single LinkedIn caption based on the provided blog post, narrowing the task to one concise outcome
Tone Conditioning
The AI is directed to begin with a thought provoking hook and maintain a confident professional tone throughout the caption.
Content Enhancement
The prompt advises not to include links and to use relevant hashtags to improve discoverability without breaking LinkedIn best practices.
Output Formatting
The AI is told to return only the caption with no formatting, markdown, or extra commentary, ensuring the result is clean and ready to post.
Input Referencing
The prompt feeds the blog content using Blog post with the variable { $('Blog Creation').item.json.message.content }} to anchor the AI’s response contextually

You are a Professional Linkedin Content Strategist

X (Formerly Twitter) Caption

Smart Prompting

Your prompt looks like:

You are a digital strategist writing for X (formerly Twitter).  

Summarize the blog post below into a single, punchy tweet.  

Focus on one key insight or hook. Keep it under 280 characters. Do not include links.

Use relevant hashtags and light emojis to boost engagement.  

Return only the tweet—no formatting, markdown, or commentary. 

Blog post: {{ $('Blog Creation').item.json.message.content }}

Prompt Breakdown

Prompting Technique

How it's used?

Role Instructioning
The AI is positioned as a digital strategist writing for X which frames the tone and format expectations for the tweet.
Task Specification
The prompt clearly instructs the AI to summarize the blog into a single tweet giving it a well defined objective.
Tone Conditioning
The AI is directed to focus on one key insight or hook keep it under 280 characters and maintain a punchy tone suitable for X.
Content Enhancement
The AI is instructed to avoid links and to use relevant hashtags and light emojis helping to improve visibility and engagement.
Output Formatting
The AI is asked to return only the tweet with no formatting markdown or commentary ensuring a clean concise social media output.
Input Referencing
The prompt provides the blog content through Blog post using the variable {{ $('Blog Creation').item.json.message.content }} to ground the tweet in real input.

4Generate High-Impact Email Copy with OpenAI

To complete the content workflow, the agent generates a short, conversion focused email derived directly from the blog post. The AI is instructed to act as a professional email copywriter and follows a structured prompt: open with a teaser to hook the reader, follow with a concise value summary, then a clear call to action, and end with a warm, friendly closing.

The output excludes formatting, or metadata, delivering only clean, ready to send copy. As with social captions, prompt engineering plays a central role in defining tone, structure, and clarity, enabling the client to scale personalized email content without manual drafting or revision.

You are a friendly, conversational email copywriter.

Email Copy: Smart Prompting

Your prompt looks like:

Prompt Breakdown

Prompting Technique

How it's used?

Role Instructioning
The AI is given the role of a friendly conversational email copywriter aligning its tone and style to suit warm approachable messaging.
Task Specification
The prompt directs the AI to write a 3 to 6 sentence email based on a blog post clearly defining the expected length and content format.
Tone Conditioning
The AI is instructed to begin with a curiosity driven hook maintain a helpful brand voice and close with an inviting call to action
Output Formatting
The AI is asked to return only the email body without any subject line or preview text ensuring a clean usable output
Input Referencing
The blog content is referenced using Blog content with the variable {{ $('Blog Creation').item.json.message.content }} anchoring the response in source material.
Content Injection
The prompt includes an instruction to add a specific blog link at the end of the email ensuring the output includes a clickable call to action destination

You are an expert email copywriter.

Email Subject: Smart Prompting

Your prompt looks like:

You are an expert email copywriter.

Based on the content below, generate a compelling subject line that captures attention and summarizes the message.

Respond with only the subject line—no explanation or extra text.

Email content: {{ $json.message.content }}

Prompt Breakdown

Prompting Technique

How it's used?

Role Instructioning
The AI is assigned the role of an expert email copywriter guiding it to write subject lines with marketing clarity and professional tone.
Task Specification
The AI is instructed to create a compelling subject line based on the email content focusing on attention capture and summary.
Tone Conditioning
The AI is told to respond with only the subject line and to exclude any explanation or extra text ensuring a clean one line result.
Input Referencing
The prompt includes Email content using the variable {{ $json.message.content }} as the input source giving the AI clear context for generating output.

5Recording AI-Generated Data to Google Sheets

All content outputs including blogs, social media captions, email subjects, and email bodies are stored directly in Google Sheets, linked to their corresponding topic entries.

This centralized setup keeps everything organized, easy to access, and ready for review. Clients can quickly update, track progress, or repurpose content without switching between tools, making the entire process more efficient and user friendly.

6Publishing Social Posts

to Facebook, LinkedIn, and X

Using n8n, the generated captions for Facebook, LinkedIn, and X are automatically published through native or API-based connections to each platform. Each post is sent with its platform-specific tone and structure, ensuring consistent and optimized messaging.

The process runs fully automated, eliminating the need for manual scheduling tools and enabling the client to maintain an active multi-channel presence with zero effort.

7Sending Emails to Clients or Subscribers

Using n8n, the generated email content is automatically sent to the intended recipients. The email body stored in Google Sheets is pulled into the email step and delivered as plain text, ensuring consistency and clarity. This step runs without manual intervention, allowing the client to maintain timely communication with their audience while keeping the workflow simple and fully automated.

8Monitoring Workflow to Eliminate Redundancy

After the blog, social media posts, and emails are published or sent, the agent updates the corresponding Google Sheet entries to mark them as checked. This prevents duplication, keeps the content dashboard organized, and ensures that each topic is processed only once. The automation maintains a clear workflow history, helping the client stay on top of all scheduled outputs without manual tracking.
Results at a glance.
This wasn’t just automation. It was a complete transformation. By combining prompt design with low code automation, the client activated a hands free content engine that delivers consistent, on brand output. Here’s what changed:
80 Percent Time Reduction
The agent now produces and schedules four full content sets each month, freeing the client from manual writing, formatting, and posting.
Consistent Brand Voice
Role based prompts and tone conditioning ensured every blog, caption, and email reflected the client’s brand, solving the mismatch issues that caused delays with previous contractors.
Fully Automated Workflow
Make.com powers the entire process, from Airtable input to Buffer scheduling. The client never needs to open a content tool.
Everything in One Place
All content is saved in Airtable, organized by topic and linked across formats. No more scattered files or lost versions.
Duplication Control
Once content is published, the related Airtable record is marked as “Checked,” preventing repeat outputs and keeping the calendar clean.
Scalable by Design
The prompt framework is modular, making it easy to increase volume whether it's four posts a month or thirty without changing tools or adding team members.

Strategic and Prompt Engineering Insight

This layered structure, moving from topic to blog, blog to captions, and blog to email, is powered by prompt chaining and role specific instruction. Each step builds on the last, preserving tone, structure, and message clarity while eliminating the need for manual editing or formatting.

All of this runs inside Make.com, so the client never needs to manage prompts, models, or tools. They simply add a topic and receive fully written, platform ready content with no friction and no extra work.

The result is a lean and reusable content engine that delivers high quality, brand aligned content at scale, helping the client stay consistent without increasing effort.
This implementation proves a critical insight: the real engine behind scalable AI automation is not just the model. It is the design of the prompts that guide it. While GPT 4o powered the content, the results came from how each prompt was intentionally crafted, connected, and aligned across every step of the workflow.

The process follows a clear and modular pipeline. It begins with a single input, a curated topic pulled from Airtable. That topic is passed into a zero shot prompt that generates a full length blog post written in a consistent and professional brand voice. This blog then becomes the foundation for platform specific captions, each created by a tailored prompt written for Facebook, LinkedIn, and X. Finally, the same content is used in an email prompt that produces a warm and engaging message with a clear call to action.

What's next?

Book a Call

This AI agent shows that when automation is guided by smart prompt design, the result is not just efficiency. It is transformation.

If you are tired of juggling disconnected tools, rewriting the same content, or struggling to stay consistent with your brand, this is your blueprint.

Built once. Scales without limits.

Whether you are a solo founder, a lean team, or a growing agency, this automation agent can be cloned, tailored, and launched to match your exact workflow and voice.

Ready to build

your own AI-Powered content engine? Ready to automate your content without sacrificing quality?

Fill out the form and let’s get started.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.