In the world of digital marketing, content is the currency of attention. One winning slogan or post can bring in hundreds of customers, while a poor text can waste your ad budget. But how do you create content quickly, consistently, and in the right tone?

In this lesson you will learn to:

• create effective prompts for content generation;
• configure the “brand voice” through a prompt;
• automate marketing sequences — from the first email to the follow-up.

We will use ChatGPT and TypingMind as tools for automation and creativity.

📚 What You’ll Learn

• How to formulate prompts for generating posts, emails, and landing pages
• How to set a brand’s style, voice, and tone through a prompt
• How to build email sequences: welcome, reminder, follow-up
• How to measure engagement and A/B-test variants

📘 Key Terms

• Brand voice — the combination of tone, style, and vocabulary in which a brand “speaks” to its audience. Example: friendly and conversational — like Duolingo, or formal and expert — like McKinsey.
• Follow-up — an additional message after the first contact. Used for reminding, clarifying, and retaining the client. Common in email campaigns and sales funnels.
• CTA (call to action) — an appeal that encourages the user to perform a desired action: “Buy,” “Subscribe,” “Learn more.” One of the most important elements in copy.
• AIDA — a classic model for structuring marketing copy:
  • Attention — capture attention,
  • Interest — spark interest,
  • Desire — increase desire,
  • Action — prompt action.
• TypingMind — a layer on top of ChatGPT that lets you save prompts, use templates, track history, and conveniently organize workflows. Useful for marketers and SMM specialists.

✍️ Prompts for Content Generation

Text generation is one of the most popular and effective applications of LLMs. But just saying “write a post” is not enough. Like a human, the model needs context: for whom, where, and in what style to write. The more precise the prompt, the more precise the result.

1. Social Media Posts

📌 Sample request:

Write an Instagram post for a sustainable clothing brand. Topic: “How to choose pieces that last.” Tone: friendly, informative. Add a CTA at the end.

📌 What to pay attention to:

• Platform matters: Instagram favors emotion, storytelling, and emojis; LinkedIn — facts and a professional style; Telegram — a direct and conversational tone.
• Target audience — like choosing a character in a game. You’ll tell a teen, an entrepreneur, and a same-age parent different things using the same words.
• Constraints and style — like a technical brief for a copywriter:
  • length (e.g., up to 280 characters for Twitter),
  • presence of emojis or hashtags,
  • style (friendly, formal, bold),
  • format (list, storytelling, Q&A).

📌 Analogy: Writing a post without a prompt is like telling an artist “draw something.” A precise prompt is like a sketch and color palette: the result comes in the right style, for the right people, with the right effect.

2. Headlines and Slogans

📌 Sample request:

Come up with 5 slogans for a food delivery service. Emphasis: speed and freshness. Style: bold, memorable.

📌 How to strengthen the result:

• Ask for rhyme, wordplay, metaphors: → “Hot. Fast. At your door.” → “By the time the kettle boils — dinner’s en route.”
• Specify the desired vibe: bold, cozy, refined, B2B — it’s like choosing a poster font.
• Specify length: short, one beat, no more than 5 words.

📌 Analogy: Crafting a slogan is like “Name That Tune” with three notes. You must evoke emotion in 3–5 words. LLMs do well if you set a clear style and goal.

3. Landing Page Copy

📌 Sample request:

Write landing page copy for an online design course. Target audience: beginners who are afraid to start. Use AIDA: Attention → Interest → Desire → Action.

📌 What’s important to specify:

• TA: newbies, pros, people with a pain point — the approach differs.
• Fears and doubts: like a boss’s weak spots in a game — if you name them, the text hits the mark.
• Structure: you can ask for AIDA, PAS, StoryBrand, or simply:
  • Headline
  • Subheadline
  • Benefits
  • Social proof (reviews, numbers)
  • Call to action

📌 Analogy: Without structure, a landing page is like a room without furniture. Pretty, but unclear what goes where. With AIDA, the model doesn’t just “write text,” it builds a persuasion script step by step.

🗣 Setting the Brand Voice

The same information can sound completely different depending on tone and delivery style. That’s brand voice: a set of words, intonations, and images that make it feel like a person is talking to you, not a text generator.

🧩 Why it matters

Without setting the voice, all ChatGPT texts will sound the same: polite, slightly academic, and faceless. In the real world, people react to character, emotion, and distinctiveness. As in ads: people remember not the fact, but the intonation and style.

📌 Analogy: Imagine you’re at a café and a robot reads the menu in a monotone. Now imagine a charismatic barista who says:

“We’ve got mocha — like a cozy Saturday in a cup. And espresso — a slap of wakefulness.”

The difference isn’t in content but in voice.

🛠 How to set brand voice in prompts

Add an instruction directly in the prompt:

“Write like a brand with personality: confident yet friendly. Use metaphors, avoid clichés. Don’t be afraid to be bold.”

🔑 Possible parameters:

• Confidence level: confident vs modest
• Style: ironic vs formal
• Lexicon: simple vs professional
• Imagery: with metaphors vs literal

📌 Tip: Specify which person the brand resembles. This helps the model maintain the style.

🧭 Sample voice archetypes

📌 Analogy: You’re the director; the LLM is the actor. The same script can be played as a comedy, drama, or detective story. Your job is to set the task and define the style of performance.

💡 Consider how structure, wording, and attitude toward the reader change. Try taking a fragment and rewriting it in another tone — a great technique for tuning prompts.

📧 Automating Sequences

An automated email sequence is like a polite, caring sales funnel, just in email format. The main thing is not to “pressure” the user but to guide them: gently, logically, step by step.

🧩 What is a sequence?

A sequence is a series of emails sent automatically based on time or user actions (e.g., opened an email or clicked a link).

📌 Analogy: Imagine you walk into a bookstore.

• First, a consultant greets you (Welcome).
• Then they gently remind you about a discount on a bestseller (Reminder).
• Then they share other readers’ reviews (Follow-up). You feel cared for — and make a decision consciously, not under pressure.

✉️ Example of a basic sequence

1. Welcome email

📍 When: immediately after registration

📍 Goal: greet, explain value, give a helpful link

Prompt:

“Write an email on behalf of an online Python learning platform. Greet the user, briefly explain what’s ahead, and suggest starting with a simple intro lesson. Up to 150 words, friendly tone.”

2. Reminder

📍 When: after 2 days

📍 Goal: regain attention + motivate via benefits

Prompt:

“Create a second email if the user hasn’t started learning. Remind them about the course, emphasize that it’s free, and suggest starting with a short video. End with a call to action (CTA).”

3. Follow-up

📍 When: after 5 days

📍 Goal: build trust via a testimonial

Prompt:

“Compose an email with a student story. Note how they were afraid to start but completed half the course in 2 weeks. End with a registration link.”

🛠 How to automate with TypingMind

TypingMind — a convenient interface over ChatGPT — allows you to:

• Save prompt templates for all emails;
• Record sequence history — see which step performed best;
• A/B test email versions;
• Track click and open analytics.

📌 Analogy: Like a CRM in sales: you’re not just sending emails — you’re building relationships, seeing where a person “gets stuck,” and picking the right text at the right moment.

✅ Tips for sequences

• Email length — up to 150 words. Better 80.
• One CTA at the end — don’t overload.
• Change tone by step: Welcome — warm, Reminder — confident, Follow-up — inspiring.
• Use the brand archetype (see previous section).

🧨 Common Mistakes

• Overcomplicated prompts with no focus: specify format and length.
• Mixed tones: post starts expert and ends with a meme.
• Missing CTA: the model won’t always add one by default.

❓ FAQs

Q: Can the model write like a specific brand?

A: Yes. Provide 2–3 sample texts — and ask it to mimic the style.

Q: Can I automate an entire email sequence?

A: Yes. Ideally via templates in TypingMind or GPTs with “Memory.”

Q: How do I track which prompt is more effective?

A: Through A/B testing: different emails to different groups, then compare opens and clicks.

📌 Summary

Today you learned how to:

• formulate prompts for marketing content;
• set the brand voice in generation;
• build automated email sequences;
• avoid common mistakes and test hypotheses.

Next, we’ll look at prompts for code generation and data analysis — they require a different structure, but the same evaluation principles apply.