
In the first and second parts of conversion copywriter Chris Silvestri’s series on empathy engineering—his framework for simulating customers using AI—he wrote about the process of gathering and analyzing data and making AI customer personas. In the third part of this four-part series, he starts communicating with his virtual creations, testing his marketing messages against their different personas.—Kate Lee
Was this newsletter forwarded to you? Sign up to get it in your inbox.
What if you had a room full of your ideal customers, eager to give you instant feedback—without the cost and hassle of a real focus group? Artificial intelligence can help.
In the first two parts of this series, we laid the groundwork for empathy engineering, a framework for getting inside your customers’ heads by simulating them with AI. We learned how to gather rich customer data from various sources and then turned those insights into detailed AI personas. Now, it's time to put those personas to work, using them as our virtual sounding boards, testing our marketing copy, and uncovering hidden opportunities for improvement.
But we can never expect AI to do all the heavy lifting for us. We have to work with it, and the results are worth it. We need to instruct, educate, and provide context to our large language model, so it can fill in the gaps in our understanding and help our messaging truly resonate with our target audience.
We’ll use AI to generate sample copy, which we'll then test and refine, so you can see how this process works. But before we do that, we need to understand how buying decisions happen.
Stepping into your buyer’s journey
We don't make decisions in a vacuum.
We go through a process: We become aware of a problem, consider our options, weigh the pros and cons, and then—finally—make a choice. Our feelings and past experiences influence each step along the way.
It’s the same when your prospects make buying decisions online. That’s why, by guiding our AI personas through the different stages of the customer journey, we can help them process information more like a human by considering the context, emotions, and shifting priorities that shape real-world decisions. Research shows that AIs “learn” best when they’re given information step-by-step, just like humans do. In our case, this will lead to more accurate, nuanced, and insightful responses—and ultimately, to better marketing and copywriting.
Customers go through emotional and psychological journeys every time they buy something. This is where the "transformation message map" comes in.
Created by conversion copywriter Joanna Wiebe, this framework helps us visualize how the customer feels when they first encounter a problem (the early stage), what they dream of achieving (the dream state), what concerns they have before making a purchase (the consideration stage), and how they envision using the product once they've bought it (the decision stage).
We tested this framework by prompting the AI persona we generated for TeamFlow–our made-up AI task management company from parts one and two–to go through a hypothetical buying journey and articulate their thought process at key stages.
Here are the results on our first ideal customer persona (ICP), named Jane, who is an “efficiency-driven project manager”:
Source: Screenshots from Google Gemini 1.5 Pro.The screenshots show Jane's responses to each stage of the transformation message map. Notice how her motivations shift as she moves from initial awareness to becoming a paying customer. In the early stage, she's primarily focused on solving immediate pain points, like overwhelm and high stress within her team, but by the decision stage, she's envisioning the long-term benefits of using TeamFlow, like an increase in productivity and the potential to take on more clients. This map will serve as a guide as we test our marketing copy, ensuring that our messaging aligns with the customer's mindset at each stage. You’ll want to repeat the same process with all of your AI personas.
TeamFlow’s copy: Before and after
For this exercise, we'll be working with two versions of TeamFlow's homepage. The first is an example of what not to do: It's generic, uninspired, and misses the mark when it comes to connecting with TeamFlow's target audience (you can view the site along with reasons why each section doesn’t work). Think of it as the company’s current homepage copy that we need to improve.
Source: Courtesy of the author.How can we create new and improved homepage copy so we can test it against the current version? By considering positioning—how a product or brand should be perceived by its target audience relative to competitors—we can write more effective and differentiated messaging.
In your positioning strategy, you should answer three questions:
- What does your company do?
- Who can your product help?
- How do you do it better or differently than competitors?
For the purpose of our project, here’s what we know about TeamFlow:
- Target audience: Efficiency-seeking project managers
- Value proposition: AI-powered task management that simplifies workflows, reduces stress, and improves team collaboration
- Key differentiators: Seamless integrations, user-centric design, and AI-driven task optimization
In the first and second parts of conversion copywriter Chris Silvestri’s series on empathy engineering—his framework for simulating customers using AI—he wrote about the process of gathering and analyzing data and making AI customer personas. In the third part of this four-part series, he starts communicating with his virtual creations, testing his marketing messages against their different personas.—Kate Lee
Was this newsletter forwarded to you? Sign up to get it in your inbox.
What if you had a room full of your ideal customers, eager to give you instant feedback—without the cost and hassle of a real focus group? Artificial intelligence can help.
In the first two parts of this series, we laid the groundwork for empathy engineering, a framework for getting inside your customers’ heads by simulating them with AI. We learned how to gather rich customer data from various sources and then turned those insights into detailed AI personas. Now, it's time to put those personas to work, using them as our virtual sounding boards, testing our marketing copy, and uncovering hidden opportunities for improvement.
But we can never expect AI to do all the heavy lifting for us. We have to work with it, and the results are worth it. We need to instruct, educate, and provide context to our large language model, so it can fill in the gaps in our understanding and help our messaging truly resonate with our target audience.
We’ll use AI to generate sample copy, which we'll then test and refine, so you can see how this process works. But before we do that, we need to understand how buying decisions happen.
Sponsored by: Every
Tools for a new generation of builders
When you write a lot about AI like we do, it’s hard not to see opportunities. We build tools for our team to become faster and better. When they work well, we bring them to our readers, too. We have a hunch: If you like reading Every, you’ll like what we’ve made.
Stepping into your buyer’s journey
We don't make decisions in a vacuum.
We go through a process: We become aware of a problem, consider our options, weigh the pros and cons, and then—finally—make a choice. Our feelings and past experiences influence each step along the way.
It’s the same when your prospects make buying decisions online. That’s why, by guiding our AI personas through the different stages of the customer journey, we can help them process information more like a human by considering the context, emotions, and shifting priorities that shape real-world decisions. Research shows that AIs “learn” best when they’re given information step-by-step, just like humans do. In our case, this will lead to more accurate, nuanced, and insightful responses—and ultimately, to better marketing and copywriting.
Customers go through emotional and psychological journeys every time they buy something. This is where the "transformation message map" comes in.
Created by conversion copywriter Joanna Wiebe, this framework helps us visualize how the customer feels when they first encounter a problem (the early stage), what they dream of achieving (the dream state), what concerns they have before making a purchase (the consideration stage), and how they envision using the product once they've bought it (the decision stage).
We tested this framework by prompting the AI persona we generated for TeamFlow–our made-up AI task management company from parts one and two–to go through a hypothetical buying journey and articulate their thought process at key stages.
Here are the results on our first ideal customer persona (ICP), named Jane, who is an “efficiency-driven project manager”:
Source: Screenshots from Google Gemini 1.5 Pro.The screenshots show Jane's responses to each stage of the transformation message map. Notice how her motivations shift as she moves from initial awareness to becoming a paying customer. In the early stage, she's primarily focused on solving immediate pain points, like overwhelm and high stress within her team, but by the decision stage, she's envisioning the long-term benefits of using TeamFlow, like an increase in productivity and the potential to take on more clients. This map will serve as a guide as we test our marketing copy, ensuring that our messaging aligns with the customer's mindset at each stage. You’ll want to repeat the same process with all of your AI personas.
TeamFlow’s copy: Before and after
For this exercise, we'll be working with two versions of TeamFlow's homepage. The first is an example of what not to do: It's generic, uninspired, and misses the mark when it comes to connecting with TeamFlow's target audience (you can view the site along with reasons why each section doesn’t work). Think of it as the company’s current homepage copy that we need to improve.
Source: Courtesy of the author.How can we create new and improved homepage copy so we can test it against the current version? By considering positioning—how a product or brand should be perceived by its target audience relative to competitors—we can write more effective and differentiated messaging.
In your positioning strategy, you should answer three questions:
- What does your company do?
- Who can your product help?
- How do you do it better or differently than competitors?
For the purpose of our project, here’s what we know about TeamFlow:
- Target audience: Efficiency-seeking project managers
- Value proposition: AI-powered task management that simplifies workflows, reduces stress, and improves team collaboration
- Key differentiators: Seamless integrations, user-centric design, and AI-driven task optimization
With these key points in mind, here’s version B of TeamFlow’s homepage copy, divided into its core sections and subsections along with improvement notes based on the positioning strategy (view it in full).
We'll compare our two homepage versions to see which one resonates more strongly (aka A/B testing) with our efficiency-driven project manager, Jane.How to A/B test your copy with AI personas
To test two versions of the homepage, we'll use a structured prompt that guides our AI persona, Jane, through a simulated experience of her visiting TeamFlow's website.
In this prompt, we instruct our AI persona to compare two versions of product copy by evaluating each against specific criteria and answering related questions. We want to know how well each copy variant:
- resonates in language and tone,
- addresses unspoken concerns,
- validates the prospect’s needs,
- frames the solution clearly,
- anticipates future needs,
- maintains relatability, and
- empowers the reader’s decision-making.
Then we ask it to give a score for each of these criteria. The goal is to refine our messaging by identifying the version that best aligns emotionally and cognitively with our target audience.
With Jane in mind, the AI analyzed both versions of the copy and provided detailed feedback:
Source: Screenshots from here forward from Google Gemini 1.5 Pro.By analyzing Jane's feedback, we can pinpoint areas for improvement, identify hidden objections, and refine our messaging to create a more compelling and persuasive experience for our target audience. Our AI persona is clearly telling us where version A is falling short: It doesn't get her attention or address the nuances of her specific pain points. Then she explains why version B does a better job at persuading her: It feels “exciting” because it hits her exact pain points and connects them with product solutions, like AI automation and workflow optimization. Version B—the one informed by the empathy engineering process and research—emerged as the clear winner.
(To make this test as “objective” as possible, I used a separate chat session with an entirely different LLM [ChatGPT-4o] to write the copy for Version B. This ensured that the AI evaluating the copy [running on Google Gemini 1.5] was not influenced by any previous interactions or knowledge of the "winning" version.)
Using AI personas as your thought partners
While A/B testing is valuable, empathy engineering goes beyond simple comparisons. With feedback from your AI persona, you can explore ideas, uncover hidden opportunities, and refine your messaging with a level of depth and nuance that traditional research and testing methods often miss. And you can do it faster.
Here are a few examples of how I use AI personas to refine my copy:
1. Probing for understanding and emotional resonance
Your AI persona might tell you that it "gets" your copy, but we want to peel through the layers and find out in what way it does so. Does it trigger the right emotions? Does it connect to their core pain points? Does it use language that feels authentic and relatable?
Here's an example prompt I used with Jane to gauge her emotional response to our message in a potential headline for TeamFlow:
Based on our goals, we could use a similar prompt to understand whether or not our copy is clear or confusing. For example, by asking: “Does this make sense to you? Is there a clearer way to say it”?By analyzing Jane's response, I can evaluate her understanding of the headline's message, her emotional reaction, and whether the language resonates with her experience. Jane seems to click with our copy—she told us that getting her time back evokes hope, that she’s looking for “intelligent” solutions, and that the copy about remote teams feels relatable. But what if she’d told us that the term “AI-powered” sounded too “buzzword-y”? Then we might have adjusted our message accordingly and tested it again.
2. Uncovering unspoken needs and desires
Sometimes, the most valuable insights are the ones we don't expect. AI personas can help us uncover those hidden needs and desires that our customers might not even articulate themselves.
Here's an example prompt where we ask Jane to imagine a near future where she’s been using the tool successfully. We want to learn about what would surprise her and surpass her expectations:
Jane's response might reveal customer needs or desires that we haven't explicitly addressed in our copy, leading to new messaging angles. For example, we learn that she was worried about how quickly her team would adopt the new tool, but that it has been surprisingly quick and painless. She credited the intuitive interface and integrations, two aspects we could highlight more prominently in our messaging, if we haven’t yet. Jane also tells us that the tool’s ability to automatically manage team workloads proactively exceeded her expectations (even factoring in upcoming vacations, which is a golden nugget to use in our copy!). I’d definitely want to emphasize this point in a headline.While the AI can't evaluate features it hasn't used, it can tell us which features are perceived as most valuable or desirable based on the research data it does have, which can help with prioritization in our product roadmap.
3. Simulating real-world scenarios
One of the most powerful applications of AI personas is role-playing in real-world scenarios, allowing us to step into the customer's shoes and see how they might interact with our product or brand in different contexts.
Let’s say we want to get a deeper understanding of the internal decision-making dynamics for a company like Jane’s. We can prompt our AI persona to picture itself in a specific situation (like bringing up the case for TeamFlow to her CEO) and tell us what she would think or say:
By analyzing Jane's response, we can gain insights into:- Key benefits: What aspects of TeamFlow does she find most valuable?
- Potential objections: What concerns does she anticipate from her CEO?
- Persuasive arguments: How does she address those objections and make the case for TeamFlow?
For instance, we see that Jane values AI automation, improved team collaboration, and data-driven reporting. We also learn that her CEO will likely be worried about how all of these provide a demonstrable return on investment that they might initially see the tool as a cost rather than as an investment. It’s also helpful to look at how Jane structures her argument, starting with her team’s three key challenges (in a specific order, too).
With empathy engineering, we don’t just want to make our marketing faster or cheaper; our goal is to also better connect with our customers. What you learn might not be the absolute truth (there’s always bias and nuance in human decision-making), but when you combine what you learn with your own creativity and expertise, you’ll gain a level of insight that was previously impossible.
Empathy engineering isn’t over yet
We've mapped the customer journey, crafted compelling copy, and started testing our messaging with AI personas. But we’re never really done. Empathy engineering involves continuous refinement, adaptation, and tweaking to ensure we stay aligned with the ever-changing reality of our customers.
In the final piece of this series, we'll go through “harmonization.” We’ll refine our prompts, adjust our AI personas, and create a powerful, iterative cycle between customer feedback and our marketing.
Chris Silvestri is the founder of Conversion Alchemy and a conversion copywriter for B2B SaaS companies.
To read more essays like this, subscribe to Every, and follow us on X at @every and on LinkedIn.
We also build AI tools for readers like you. Automate repeat writing with Spiral. Organize files automatically with Sparkle. Write something great with Lex.
Ideas and Apps to
Thrive in the AI Age
The essential toolkit for those shaping the future
"This might be the best value you
can get from an AI subscription."
- Jay S.
Join 100,000+ leaders, builders, and innovators

Email address
Already have an account? Sign in
What is included in a subscription?
Daily insights from AI pioneers + early access to powerful AI tools
Ideas and Apps to
Thrive in the AI Age
The essential toolkit for those shaping the future
"This might be the best value you
can get from an AI subscription."
- Jay S.
Join 100,000+ leaders, builders, and innovators

Email address
Already have an account? Sign in
What is included in a subscription?
Daily insights from AI pioneers + early access to powerful AI tools