How to Create Laser-Focused Target Audience Personas with AI

Full video transcript

How’s it going, all? Darby here with Gen AI University, and in this video you’re going to be walked through a 90-minute pre-live workshop that we just wrapped up. At the beginning of the workshop we actually had a bit of technical difficulty with ChatGPT going down, so we did a quick pivot over to Claude with Anthropic and ran the entire demonstration of part one, the Buyer’s Blueprint, from the Buyer’s Brief book, creating an entire detailed blueprint that ended up being about 30 pages of specific audience-focused language and messaging that we could take and run within a campaign.

The purpose of this workshop is to help give you the workflow and the skill insights that you can literally leverage in your business to create these personas for yourself, for your clients, and demonstrate the value of any particular brand as it relates to the audience and the market that you’re going to be positioning your product and service to. You might find new markets and segments that you wouldn’t have considered before, or create more refined, laser-focused versions of your existing audience by implementing the workflow that we’re demonstrating in this class with the Market Sauce Method.

So go ahead and check out the links below this video, feel free to pause, run the workflow side by side. It would be helpful to have a copy of our book, the Buyer Brief book, but you can also use the free Market Sauce Mentalist GPT that’s in the description below to run alongside how we’re demonstrating it. Again, this applies to Claude, ChatGPT, and a number of other large language tools and models that you may use in your business.

With that said, over the next 90 minutes it’s going to be an introduction to the workshop, and then we’re going to have our deep dive walkthrough with the demonstration, showing in action how these tools will interact with each other. I hope you find it useful and actionable for yourself with the free tools that we’re providing, and if you like the content we produce, please subscribe to our channel and like this video and share it with friends who you think would find it valuable.

All right everybody, good morning and welcome to today’s free live workshop that we are hosting on creating customer personas with AI, creating laser-targeted customer personas to focus in on our ideal target audiences and creating more engaging, more relevant, more effective content. Great to see everybody here today tuning in with us live, and if you’re catching the recording on YouTube, great to see you too.

We got 90 minutes blocked off for today’s workshop and we’ll be walking through a very specific workflow for identifying core audience demographics. We’re also going to be covering psychographics, market segmentation, and how to craft messaging that speaks directly to these different segments of our audiences. The focus today is going to be very strictly on persona development and market segmentation. We will be covering additional things that you do once you identify the segments and create additional campaigns along that line.

Part of working with AI is knowing how to adapt and iterate on your process whenever these tools do things — AI is going to AI, as we say. So what we’re going to do is start with the introduction to the presentation, leave a little more context for you joining us here, and then dive into the actual demonstrations. While we’re going through this workshop there is an interactive element to it, where you’re going to have access to prompts to hopefully GPTs that’ll work for you here while you’re going through this live and inputting information and context on your end to start to develop these personas on your own.

We’re going to be talking about creating target audience personas with AI, and the purpose of this is to help you with revolutionizing your own process and leveraging AI and the workflows that we’re going to be sharing with you today to really get the most out of this technology. Whether we’re demonstrating ChatGPT or Claude in this case, or if you use another tool, there’s a very collaborative process the way that I work with these tools, and I hope you’ll find some value to use in your own approach as we’ve developed at the Market Sauce Method.

The objective of this workshop is, first and foremost, creating personas and making sure these personas you’re creating are accurate, using AI to generate precise audience profiles from ideally your own existing content and customer interview transcripts — we’ll also be doing it just from scratch so you can see how the workflow works both ways. It’s also about boosting your efficiency during this process. What we’ve outlined inside of our Buyer Brief book, with the Buyer Blueprint section of that book, might take a professional copywriter three to four hours of research time to go and manually find, research, interview, and put together a blueprint. That is essentially the persona you’re going to be creating a marketing campaign around. If you’re not a professional marketer, that process could take you weeks, or more likely never get done, because it’s a very time-intensive manual process.

Creating super effective marketing personas is going to help you with directly improving your bottom line in a number of different ways, and ultimately using AI in the way we’re demonstrating here helps you uncover new opportunities and the language and messaging to speak to each one of those different personas and segments of the market, which will open up new innovations on your own products or new ways you can create and improve on what you’re already delivering.

In the demonstration today I will be primarily using ChatGPT, for which we have a custom GPT that we built, and I’ll also demonstrate how you can approach this process inside of Claude. Both of the ones I mentioned require a paid membership — I suggest if you don’t already have a ChatGPT-4 membership, that’s about 20 bucks a month, I think Claude is similar in pricing. These are both incredibly valuable tools, staples inside of my workflow, so I’d recommend integrating at least one of these into this process if you aren’t already.

We’re also going to be demonstrating how to use some of the prompts within the Buyer Brief book inside of these tools, and showing you how to improve, expand, and build upon these prompts and your own instructions to guide the AI in the direction you’re looking to accomplish. Whether you’re doing B2B marketing outreach campaigns or creating direct-to-consumer ad campaigns for an e-commerce brand, there’s a number of ways to interact and engage with your persona once you’ve got the basic foundation made. That’s where the Market Sauce Mentalist GPT, as a free resource to you, comes into play — you can leverage this GPT and the customer insights you bring to the table to create more effective, targeted customer personas that become the foundation of any go-to-market campaign you’re creating.

The method we take from Gen AI University all the way down to Market Sauce, and what we’re developing as a market research tool, all comes down to co-creation — how AI can be a collaborative partner inside the creation process, but keeping in mind you are the human in charge, you’re the boss, you’re guiding AI toward the goal, and setting that goal is critical. This is all about conversational interaction and developing the skill of prompting, though it’s more than just knowing how to prompt — a prompt is really an instruction using natural language to help bridge where you’re at now to where you want to be. AI right now is an incredibly effective prompt co-creator that you can use when creating prompts you see inside more elaborate recipes and workflows.

Because this is a co-created, collaborative process, we want to create but also test and expand on personas — this is very much iterative. Even if you started from scratch today with your very first persona blueprint and didn’t have any customer interviews or survey data, it’s critical to go and validate and reflect, making sure that in the real world we’re combining the AI insights we’re getting with real human experience. AI is great as an ideation and brainstorming partner, but if you just let AI take the wheel and run, you might end up a thousand miles ahead in a direction — but is it the right direction? Are you validating, directing, and choosing certain checkpoints where AI needs to be checked or refined, where you need to give it more context?

What you’re going to see throughout the workflow is defining clearly your target market and how we’re using AI to help develop and refine that, but also establishing your goals with a particular campaign or with understanding insights inside the context of the Mentalist GPT and the Buyer Brief book, and all the constraints that help hone in and keep AI in line with the end goal in mind. For example, the campaign idea AI might give you if it’s identified a persona and you say you have a $5 million a month budget is going to come out a bit differently than if you say you’re super bootstrapped with $150 to work with that month.

General tips for getting the prompts going: start with the Mentalist GPT and creating a customer persona by defining the general industry and audience information you have — who you are, what you do, how you’re different, what industry you work in, context about your product, and who you serve. After running a few different persona generations, determine the types of personas you want to create based on specific goals and areas of focus for your business. A great, simple application for these generative AI tools is saying: I serve this customer, this type of market, here’s what my business does, who is my ideal target audience — and it will give you responses, possibly including an audience you never would have considered.

Once you’ve created persona profiles, creating a synthesized conclusion or even a TL;DR of what’s been created and talked about can be really helpful, especially since these reports can run 10, 20, 50-plus pages, so you frontload the insights and action items from any given thread. For larger scale needs, you’d look at working directly with OpenAI’s playground, API, Enterprise-level support, and vector storage databases — without getting too technical, that’s for making sure your own data is secure and referenced with accuracy. What we’re showing at a smaller, more simplified scale is understanding and recognizing patterns and making connections, which is key given AI’s broad knowledge, and using those connections is what leads to putting things into action and executing on campaigns.

Getting very specific about the problems you’re defining is key to co-creation — once you get specific and refined in your personas, you can go narrow and deep and create incredibly targeted messaging that speaks to a very niche audience. In a sea of online noise, the more specific you can get about who you’re helping and serving, the better your engagement results will be.

Everything you generate from AI has to be validated with real conversations with real humans. Ways to validate AI’s insights include actually surveying your customers and using those surveys to refine your blueprints, assessing your current customer list to confirm buying behavior is consistent with what you’re seeing in your business, and using customer discovery calls and interview transcripts — running them through the Mentalist GPT or the blueprint persona creation workflow lets you interact and engage with that transcript, validate pain points, and then take insights back to customers to confirm you’re on the right track. For market trends, validate through industry reports; sometimes AI will cite sources and pull a direct link, and if you’re using a tool like Perplexity, always validate and confirm the content coming back.

I’m going to go ahead and retry ChatGPT and share the Mentalist GPT we’ll use to get started with persona development… [ChatGPT remained unavailable, so the demonstration continued in Claude.] We don’t want to waste time watching the spinning wheel of death, so instead we’re going to use my tied-for-number-one favorite generative AI platform, Claude, for the demonstration. For those here live, I’m going to link a copy of the Buyer Brief book manuscript; if you’re not here live, the manuscript is available on our website and Amazon. This won’t detract from how I’m able to show you the process — we’re just not going to have the GPT to demonstrate on this call.

A number of these other leading generative AI, large language model platforms — Anthropic’s tool called Claude 3.5 — I use ChatGPT and Claude very heavily as my top two, and I’m a fan of Perplexity as well for researching facts, statistics, and things to apply into creating personas. If you’ve used ChatGPT but haven’t experienced Claude yet, use this as a way to see where Claude is at right now — these tools are evolving rapidly with new capabilities always coming out.

If you’ve got the Market Sauce Mentalist GPT, this book is literally built into the knowledge base of that GPT, so when working with it inside ChatGPT, it’s naturally referencing the context of this book, which is how it starts to analyze, assess, and create personas based on what you input. My approach today is how to take the prompts from the Buyer’s Brief and how to even start the process without these prompts — in this case using Claude for this demo, training it to act like a market research assistant to help create these customer personas, using a demonstration for a game that’s part of a company I’m involved with called Side Hustle.

If you got the book, the recipe starts on section two, but before we get to the recipe I want to open my Claude account and start a new conversation by saying hi. If you’re going to use this for research for a specific project, copy the link to the Claude thread and save it in a project or wherever you want to reference it again — that can be really helpful, since sometimes content gets buried and a week goes by and you’re not sure where your content went. I often just use a simple Google Doc for sharing information, so I’d recommend starting a document where you can copy and paste and save information for easy reference.

I use a tool called Voice In as a dictation Chrome extension — Mac and PC have voice dictation built in too — because I find it’s easier when working with AI. You can type your prompts, but AI doesn’t need perfect grammar; it can understand what you’re saying if you just talk to it and give it your instructions.

I said: I am doing customer persona research for my business and want you to act as an expert market researcher who understands the psychology and motivations of what drives my target audience to buy. From here AI still doesn’t have any context about my brand, but now it’s thinking with the hat of a market researcher, so as the brand owner I want to give it information about my product or service and tell it to analyze it to understand who the target market is based on what I’m selling.

The product is called Play Side Hustle. We spent months doing customer research and talking to different people, understanding how the game mechanics worked, and were still not sure who all the different target audiences were and how to create campaigns for them. Had we had AI back when we started, it would have saved months of research time, helping with messaging for the Kickstarter page and everything up much faster. I went to the product landing page and copied the details, then said: here are the details, I want to identify the top 10 market segment personas that make sense for this product.

It quickly identified 10 segments: young entrepreneurs, college students, remote workers, corporate professionals, gaming enthusiasts, small business owners, freelancers, career transitioners, personal finance enthusiasts, and educators and trainers. I then gave more context — the game is basically an improv version of Shark Tank meets Apples to Apples that you can play to stimulate creativity and generate ideas — and asked it to refine the personas based on that context. It came back with a second edition: startup founders, innovation teams, corporations, entrepreneurship students, improv enthusiasts, business coaches, consultants, hackathon and workshop organizers, product managers, and marketing and advertising professionals or gaming enthusiasts.

I could go down the rabbit hole with all 10 or just one, but a really important part of engaging with these large language models is knowing the focus of what you want the content to ultimately lead to — that dictates the type of content you want. Before creating full-blown campaigns for every segment, I wanted to pinpoint one specific target market and focus in with the context already provided. I chose to explore hackathon and workshop organizers, since one of the primary segments that signs up and plays the game for free is entrepreneurial students, and event organizers running entrepreneurship programs would give exposure to new entrepreneurs.

I paused Claude and scrolled to part two of the book, which is the actual recipe. Instructions and context are absolutely key for guiding AI, but until this point we didn’t know who we were targeting — it was just “Side Hustle is a great game for everyone who likes games,” which couldn’t be more broad. Now we look at messaging for a campaign targeting Side Hustle toward hackathon and workshop organizers.

I told Claude: we want to focus on hackathon and workshop organizers, referencing the Side Hustle context given above, and execute the instructions below. Claude responded with the target audience — innovative event organizers seeking tools to engage, stimulate ideation, enhance team dynamics, and add unique elements to hackathons and workshops — along with demographics, psychographics, and behavior patterns, and then the number one goal, which is key because the detail of the persona relates to who they are as a market segment, what your brand is, what value they want to get from it, and what goal they’re looking to achieve.

The goal Claude identified was creating memorable, high-impact events that foster innovation and collaboration among participants. If I’m an event organizer at hackathons and workshops wanting that, and I know our game does that, I can position our product to help them achieve that goal — a clear win for opening the conversation. It went deeper into why they’re aiming to build it, what industries could be served, where the audience is located, company size, and buying signals — like what someone searches for online (interactive workshop ideas, hackathon planning tools, creative icebreakers for events) and whether Side Hustle shows up as the answer.

Someone asked: if you had a huge disagreement with some of the outputs, how would you adjust it? Good question — sometimes Claude or any AI gives you something that doesn’t make sense. For the sake of demonstration, I reran it and it gave a similar goal: foster memorable, productive events, inspire innovation, foster meaningful connections among participants. I didn’t disagree with that direction, but to demonstrate reframing, I told Claude: I’m not sure that’s the number one goal we want — the goal of our event organizers is that they host these events at high schools with entrepreneurship programs, so I want the goal focused on that — can you give me 10 primary goal ideas to choose from?

It came back with 10 options: inspire the next generation of entrepreneurs, enhance practical business education, increase student engagement, and others. I chose “increase student engagement through gamified learning,” which has literally been an example we’ve seen for the game as a substitute for entrepreneurship classes. I told it to continue by refining the last outputs accordingly, so the report stays focused on the right number one goal.

Now that we had the refined context, I continued through the book manuscript, pasting the next section of instructions and telling Claude to reference the prior context and execute. Doing this section by section — rather than one prompt at a time as we used to do in Jasper, which took an hour or two — the long-form, one-shot prompting approach we’ve structured is designed to get the information out quickly so you can assess it and move forward, rather than spending two hours per report.

The master buyer’s blueprint segment analysis went deeper into who this person is: innovative mindsets, student-centric approach, results-driven, collaborative nature, needs, wants, desires, behaviors, and goals, tied to creating a dynamic, inspiring learning environment. It included topics (content ideas you could turn into blog posts), the primary complaint (what people are trying to move away from — for example, a stale curriculum and the frustration of watching students’ eyes glaze over during another theoretical lecture), and resources (limited budget, time crunch). Each section of the Buyer Brief book has prompts tied to it that you use in sequence.

Due to time constraints we didn’t go through every section, but as you can see, these prompts are the instruction, and seeing where they fit inside your marketing campaigns is part of the power of the Market Sauce Method — getting segments broken down and using these insights to make more strategic decisions while spending less time and resources sourcing this information. What might take hours of manual research, we got through in about 30 minutes live, and it can be done in five to 10 minutes without walking through it step by step; with our own automations, our done-for-you blueprint process is down to five minutes or less, with human quality control on top.

On negative statistics — sometimes Claude pulls links and sources, sometimes it doesn’t, and this is where you need to validate and verify. In the demo, Claude gave a statistic (only 41% of students feel like they have the skills necessary to start a business) with a cited source, but when checking the actual source, that exact number couldn’t be confirmed — a reminder that AI might get a specific number wrong, so always fact-check before citing a source in a newsletter or article.

We moved to the last two sections of the blueprint: solutions and offer target. The solutions section covered “false solutions” people try (like generic attempts to teach entrepreneurship) and how that translates into ad copy. A good way to test this is pairing pain points against ad dollars to verify which pain points perform best — using the primary and secondary complaints (the constant pressure of covering essential concepts while keeping lessons engaging, watching students’ eyes glaze over during another business talk) as the basis for ad copy framed with something like AIDA (attention, interest, desire, action) or problem-agitate-solution.

For the offer target, based on all the blueprint context, we identified additional sub-segments within this segment — high school business teachers, principals and administrators, education technology creators, curriculum development specialists — different variations of personas you could target next. From there I asked Claude to create 10 different ads in AIDA format based on the secondary complaints related to the primary goal, with a call to action to subscribe to Play Side Hustle for free at playsidehustle.com. It produced 10 ad variations, including one targeting administrative hurdles (“tired of red tape stifling your innovative teaching ideas”) tied to a pre-built, curriculum-aligned entrepreneurship program — an idea we’d actually considered, getting the tool into classrooms via curriculum.

At that point we had a 30-page detailed blueprint with ad copy ready to act on. If I need 10 quick blog post ideas later, I can upload that blueprint to a new ChatGPT thread and ask it to create outlines and newsletter ideas from that existing context, rather than starting over.

I also demonstrated the Market Sauce Mentalist GPT directly — going to marketsauce.mentalist redirects to it. It has the book built into its knowledge base, so it can assess a website and has broad context. Asking it cold to “assess the brand playsidehustle.com and create a list of top 10 personas” without more context sent it in the wrong direction, which highlights that even a knowledgeable GPT still needs the right context — I don’t want to skip providing that context just because a route seems easier. Restarting with research context — telling it to research Side Hustle on Amazon and online first — got it back on track, returning creative professionals, entrepreneurs, and startups, similar to what we found earlier with Claude. If you’ve conducted interviews and have transcripts, you can upload and leverage those transcripts to create another version of the market segment report.

Starting AI from square one with no additional data or context gets you about 80% of the way to opening conversations with clarity and confidence that you’re speaking the right language to the target market — but don’t forget to go have real conversations and validate what you’re getting, refining the blueprints over time so your campaigns become more rock solid as you gather more data and insights.

Everything starts with the persona — who you’re reaching inside your target market — creating blueprints, providing the right brand context, and using that information to guide how you get ahead of trends, assess your language, and speak directly to that target segment as you grow and expand your product line and services.

We got about 30 minutes of open floor at the end of the 90-minute block for questions. I mentioned we have a special release for our Market Sauce Revolution membership, in its early stages of rolling out, growing a community of practice inside Gen AI University for workshops, interactive implementation events, and hackathons to put these ideas into practice and develop additional workflows and resources. We’re also providing done-for-you blueprints and buyer briefs as a service, totally custom to your business, with our platform still in development for making this self-service for early adopters — typically a one-to-two business day turnaround.

I am Darby Rollins from Gen AI University, founder and CEO. This has been the Market Sauce workshop on how to create laser-focused target audience personas with AI. Check out our content and resources at genaiuniversity.com as well as marketsauce for done-for-you blueprints and briefs, and subscribe to us on YouTube for updates and more content around leveraging AI for business growth and success.