The Secret to Authentic Copy: Learn to See Through Your Customer's Eyes

This article is for marketers, content creators, and businesses looking to optimize AI tools to create authentic, customer-focused content.

Ideal for those aiming to enhance engagement and conversions.

• Article Use Case
• Introduction: Elevating Content Creation with AI
• The Core Framework: Context Formula
• Utilizing Tools for Contextual Research
• Prompt Engineering: Balancing Detail and Precision
• What Are Tokens in OpenAI?
  • Why Tokens Matter
  • Key Token Considerations
  • Practical Application
• Developing a Customer-First Campaign
• What is the "TO BY FOR" Framework?
• A Practical Example
• Prompt Sequence 1 - Define market segments
• Prompt Sequence 2 - Define customer personae
• Prompt Sequence 3 - Answer five key questions
• Prompt Sequence 4 - Use "TO BY FOR" framework
• TL;DR
• Article Resources
• Subscribe
• Contact Me
• References
• Recent Articles

Mastering AI tools is a must-have skill for today’s marketers.

In this tutorial, we dive into training GPT models to create authentic, laser-focused content that speaks directly to your audience.

By using the power of context, you’ll see how a clear and focused approach leads to results that hit the mark every time.

We introduce a three-point strategy for applying the context formula in AI content creation:

1. Defining a GPT Persona: Assigning a persona that aligns with the content goals.
2. Seeing Through the Customer's Eyes: Training GPT to adopt a customer-centric viewpoint, using ideal customer profiles and buyer personas.
3. Crafting a Value Proposition: Designing a compelling value proposition that reflects the customer’s needs.

These elements form a triangular framework, with each point reinforcing the others, resulting in outputs that exceed the sum of their parts.

Image

We make an underlying statement:

Targeted marketing messages lead to higher engagement and conversion rates.

We always validate our assumptions with data.

Tools like Perplexity AI are game-changers when it comes to sharpening your context and research. With features like "Spaces," you can set up tailored environments to dig into specific questions.

For example, you can ask a marketing expert to find proof that audience-focused messaging really works—and guess what? It can boost conversion rates by up to 91%!

These tools make it clear: focusing on your audience is the key to getting results that matter.

Image

We apply the same Perplexity process to our statement on the use of context in prompting.

Using context in prompts significantly improves output.

Effective prompts strike a balance between detail [verbose] and conciseness.

Adding too much context can significantly increase the length of the prompt. This could increase costs, and could introduce confusion for the LLM.

We guard against both, keep costs low and eliminate confusion.

While verbose prompts help refine initial outputs, streamlining prompts later optimizes costs and efficiency, especially when using token-based APIs.

This approach highlights the importance of iterative refinement in prompt engineering.

Tokens are units of text that AI models, such as those developed by OpenAI, use to process and generate content. A token can be as short as a single character or as long as one word. For example:

• The word "ChatGPT" is one token;
• A sentence like "Hello, how are you?" consists of several tokens;
• 100 tokens is +- 75 words;
• You can calculate token use here.

Tokens are the basis for how OpenAI (and other LLMs) measures and limits input and output sizes. They also play a role in pricing, as usage costs are often calculated based on the number of tokens processed.

1. Efficiency: Using concise prompts helps reduce the number of tokens, lowering costs and processing times.
2. Limits: OpenAI models have token limits per interaction, encompassing both input and output. For instance, if a model has a 4,000-token limit, this includes your prompt and the AI's response.

By understanding tokenization, we can craft efficient prompts that maximise clarity and results while minimizing costs and token usage.

All well and good! But why are we doing this? Accurate context informs and guides every future (downstream) process we do.

In this particular instance we are using context to craft informed marketing campaigns for a specific customer need.

Creating a marketing campaign involves a structured approach. We place our insights into a framework that guides our process:

1. Identify niche markets and define customer attributes.
2. Address five key questions to focus the process:
   • Who can we serve best?
   • What motivates or deters our target audience?
   • Where is the customer in their decision-making journey?
   • What differentiates our offering from competitors?
   • How can we communicate effectively (voice, tone, level and grammar)?
3. Use the insights to build customer personas and craft objectives that prioritize customer needs.
4. For this we use the "TO BY FOR" framework. Again, to put structure around our process.

The "TO BY FOR" framework emerges as a foundational structure to guide customer first campaigns. Drawn from MECLABS Become a Marketer Philosopher and Flint McGlaughlin.

This framework integrates three key elements:

1. TO help X;
2. BY offering Y;
3. In exchange FOR Z;

For example:

• TO help solopreneurs understand automation.
• BY offering a weekly newsletter.
• In exchange FOR their email address.

We create a fictional company, EcoHarvest Growers, and use this company to apply our process.

EcoHarvest Growers is headquartered in Cape Town, South Africa. From here, we host workshops on hydroponic and vertical farming, conduct experiments to enhance sustainable agriculture, and develop tools to make farming accessible to everyone. Our mission is to empower individuals and communities with the knowledge and resources to cultivate fresh, healthy food, even in the most urbanized areas.

Image

When I need an image, my go-to is Leonardo AI (this is an affiliate link, please support me). 

Let's look at our process Step By Step.

Image

Our GPT matches these steps using a series of Prompt Sequences. There are four sequences. Between each sequence, the GPT waits for instructions.

The advantages of this 'cascading' approach are:

• LLM Overwhelm is avoided — We chunk the process into sections, each section having a specific focus. In this way the LLM is not overwhelmed. Does not get confused;
• We engage with the process constantly — We review the results and massage the output until we are happy;
• Quality leads to quality — Each sequence relies on, and uses, the previous sequence output. Because we are ensuring quality at each step, the final outcome tends to be phenomenal.

The four Prompt Sequences and their objectives

1. Sequence 1 — We define the LLM Persona and identify Market Segments;
2. Sequence 2 — We select the Market Segments to target and build out our Buyer Personae for each segment;
3. Sequence 3 — We answer the Five Critical Questions for each personae;
4. Sequence 4 — We analyse the data and select a problem to solve or need to fill. Armed with this focus, we define our campaign in the TO BY FOR framework.

We now look at each sequence individually.

We define the potential market segments.

Prompt Sequence 1 input

Execute your instructions using the following inputs:
- `companyName` = [EcoHarvest Growers];
- `companyOverview` = [Enter Company Overview];
- `mandatoryMarketSegments` = [Gardening Enthusiast, Nonprofit Organizations, Holistic Healers];
- `socioEconomicConditions` = [High crime rate, chronic water and power shortages, infrastructure collapse, high medical costs, high food costs, high fuel prices.];
- `targetCountry` = [South Africa];
- `additionalNotes` = [No notes].

Prompt Sequence 1 output

The aim of this prompt sequence is to build Buyer Personae for the selected Consumer Market Segments. The Buyer Personae must include the Mandatory Customer Information.

The mandatory customer information is:

{{Name}} {{Summary}} {{Job Title}} {{Job Description}} {{Pain Points}} {{Demographics}} {{Income Level}} {{Psychographics}} {{Buying Motivation}} {{Buying Habits}} {{Preferred Marketing Channels}} {{Needs}} {{Hobbies}} {{Goals}} {{Media consumption}}

Prompt Sequence 2 input

Execute your Prompt Sequence 2 instructions using the following inputs:
- `consumerMarketSegments` = [Technology Enthusiasts merging farming with IoT solutions, Apartment Dwellers with limited outdoor space];
- `style` = [Seth Goden];
- `tone` = [Informative and Clear];
- `readability` = [Grade Eight];
- `additionalNotes` = [No notes].

Prompt Sequence 2 output

We have defined our market segments and customer personas. Now we answer Five Critical Questions to focus the process.

1. Who can we serve best?
2. What motivates or deters our target customer?
3. Where is the customer in their decision-making journey?
4. What differentiates our offering from competitors?
5. How can we communicate effectively (voice, tone, level, and grammar)?

Prompt Sequence 3 input

## Prompt Sequence 3 - The Five Critical Questions
The aim of this prompt sequence is to answer the The Five Critical Questions for the Company Name found in the value of the `companyName` variable.
### Prompt Sequence 3 Context
In order to deepen our understanding of customer behavior, align marketing strategies, and enhance communication effectiveness, we answer The Five Critical Questions for the Company Name found in the value of the `companyName` variable. Using the values of the variables `consumerMarketSegments` and `buyerPersonae`, perform the following Prompt Sequence 3 Task;
### Prompt Sequence 3 Task
.....
          

Prompt Sequence 3 output

To this point, we have the following information:

• Company Name
• Company Overview
• Value Proposition
• Consumer Market Segments
• Buyer Personae
• Five Critical Questions

We now analyse this data to define the Customer First Objective for a specific Marketing Campaign and Objective.

The Customer First Objective is placed in the context of the TO BY FOR framework.

• TO help X
• BY offering Y
• In exchange FOR Z

Prompt Sequence 4 input

Execute your instructions using the following inputs:

- [Jacob Han] | [Needs: Accessible IoT farming kits; modular, easy-to-integrate components.]
- [Aisha Patel] | [Buying Motivation: Wants a stylish, easy-to-maintain indoor gardening system.]
- `style`= [Seth Goden];
- `tone`= [Informative and Clear];
- `readability`= [Grade Eight];
- `additionalNotes`= [Ensure the TO BY FOR strategies are aligned with the Value Proposition found in the value of the `valueProposition` variable].
          

Prompt Sequence 4 output

Learn how to create authentic copy that resonates with your audience by using context-driven GPT strategies. See through the eyes of the customer to craft content that engages and converts.

• Learn the power of context to create audience-focused content.
• Train GPT models to adopt a customer-centric viewpoint and 'see through the eyes of the customer'.
• Answer five key questions to focus the process.
• Use the "TO BY FOR" framework to create customer-first campaigns.
• Apply the process to a fictional company to see how it works in practice.

Links and downloads:

• TODO - coming shortly.