{"id":1133,"date":"2025-06-25T11:00:00","date_gmt":"2025-06-25T11:00:00","guid":{"rendered":"http:\/\/buywyo.com\/?p=1133"},"modified":"2025-06-30T11:25:02","modified_gmt":"2025-06-30T11:25:02","slug":"how-to-get-ai-to-write-copy-in-your-brand-voice-from-picking-the-right-platform-to-training-your-gpt","status":"publish","type":"post","link":"http:\/\/buywyo.com\/index.php\/2025\/06\/25\/how-to-get-ai-to-write-copy-in-your-brand-voice-from-picking-the-right-platform-to-training-your-gpt\/","title":{"rendered":"How to get AI to write copy in your brand voice \u2014 from picking the right platform to training your GPT"},"content":{"rendered":"
When generative AI first came out, I was skeptical. While it came up with some good ideas, I saw it more as a novelty than a real copywriting tool. Honestly, I didn\u2019t believe \u201cgreat\u201d and \u201con-brand AI copywriting\u201d would ever belong in the same sentence.<\/p>\n<\/p>\n
To be clear, I still see my own brain and approach as a competitive advantage. That hasn\u2019t changed, but two things have <\/em>shifted:<\/em><\/p>\n And, I\u2019ve found the biggest shift by thinking of myself as a copy chief and AI as a junior writer. It has fantastic ideas but needs strong guardrails, constant feedback, and clear direction. And, I\u2019m going to show you how.<\/p>\n Table of Contents<\/strong><\/p>\n <\/a> <\/p>\n For starters, I don\u2019t believe that there is a single \u201cbest\u201d AI platform for copywriting. It really comes down to personal preference.<\/p>\n ChatGPT was the clear winner among the people I spoke with for this article, with Claude and DeepSeek close behind.<\/p>\n If you\u2019re already using HubSpot, Breeze AI<\/a> and AI Content Assistant<\/a> now includes brand voice capabilities too. The tool can help your team stay consistent across emails, blogs, and landing pages directly inside your marketing workflows.<\/p>\n Personally, I prefer ChatGPT. It\u2019s where I\u2019ve spent the most time, I\u2019m comfortable with the platform, and most importantly, I know how to engineer prompts and give feedback that gets good results. Plus, since I\u2019m already paying for it, it makes sense to focus there.<\/p>\n If you haven\u2019t played around with AI much yet, I highly recommend experimenting. Take a few different platforms, feed each the same prompts, and see what comes back.<\/p>\n Pay attention to:<\/p>\n Based on the responses, the user experience, and how well each one fits as your copywriting assistant, you can pick your go-to.<\/p>\n And remember, you don\u2019t have to commit to just one. Several copywriters and strategists I spoke with use different platforms depending on the project.<\/p>\n <\/a> <\/p>\n The first thing you need to know about training an AI platform is that you can\u2019t just say \u201cwrite like me\u201d without any context.<\/p>\n It\u2019s a garbage-in, garbage-out kind of thing, and it\u2019s on you to set it up for success.<\/p>\n So if you really want your generative AI platform to write like you, you have to teach it your voice and give it good material to work from, just like you would if you were briefing a junior copywriter.<\/p>\n Over the past few years, I\u2019ve trained ChatGPT to understand my writing style, and those of a few different clients to make sure it knows what to watch for when I need to brainstorm, draft, or edit.<\/p>\n As I was preparing to write this guide, a client project came along where we were synthesizing and updating the messaging for the parent company and its dozen sub-brands.<\/p>\n It gave me the perfect opportunity to build a clear set of guidelines for communicating across different services, topics, styles, and audiences \u2014 all while maintaining a consistent, customer-first ethos.<\/p>\n I followed the steps below, and I was blown away by how well it worked.<\/p>\n That said, I didn\u2019t want this guide to reflect only my experience.<\/p>\n You\u2019ll also find insights, quotes, and examples throughout this article from other copywriters and strategists, so you get a full, well-rounded perspective on what it really takes to train AI to write like you.<\/p>\n If you\u2019re wondering how to train AI to write in your brand voice, I want you to start by thinking of AI as a new person on your team, not just an AI platform. You have to be crystal-clear, coach it along the way, and offer constructive feedback and examples until it actually understands your voice.<\/p>\n And unlike a junior copywriter, you can tell your AI assistant that you hate what it\u2019s done, and no feelings get hurt.<\/p>\n HubSpot\u2019s Amy Marino<\/a><\/strong> shares that AI has done a good job capturing the team\u2019s customer-centric, value-focused approach to copy, whether for an internal or external project.<\/p>\n \u201cThat said, where we’ve had the most rounds of back and forth has been in my pushes to be more creative, less derivative or jargon-y, and to use concrete examples to illustrate abstract concepts,\u201d Marino says.<\/p>\n She goes on to say that sharing previous scripts, briefs, and language she loved as examples helped AI understand what she liked and wanted \u2026 and helped get to the finish line faster.<\/p>\n I\u2019ve found that the best place to start is to tell AI that you\u2019re training it on a new brand voice and name it. That way, if I say \u201cwrite like Erin Pennings,\u201d the AI I trained will know to use my voice.<\/p>\n More importantly, without this foundation, AI will default to generic copy that doesn\u2019t reflect your voice, values, or perspective.<\/p>\n I caught up with Justin Blackman<\/a><\/strong>, a brand voice strategist who has worked with tons of brands to nail their voice, beliefs, and messaging, to find out what he\u2019s seeing is working well.<\/p>\n \u201cAnyone I\u2018m doing voice finder sessions with is finding that it\u2019s not so much the copy as it is the insight that’s improved,\u201d Blackman says. \u201cWith clarification on your approach, views, and opinion, it gravitates toward the right style with fleshed-out ideas that are about 83% publishable as is.\u201d<\/p>\n Blackman says he\u2019s been through the same process he puts clients through and that he\u2019s been impressed with the output when AI is properly trained on insight<\/em> as well as tone and vocabulary.<\/p>\n Pro tip: <\/strong>As part of this step, think about spelling out the big picture:<\/p>\n Once you\u2019ve given AI the big picture, the next step is getting it to see<\/em> your voice in action.<\/p>\n This is where your real-world examples come in.<\/p>\n I like to feed AI a batch of strong samples \u2014 emails, social posts, sales pages, whatever fits. Then, ask it to analyze patterns and ask it to characterize the tone, cadence, style, and strategic choices you make without even thinking.<\/p>\n When I got in touch with my colleague and friend Chris Collins<\/a><\/strong>, a strategist who was an early adopter of AI-assisted writing, he shared that he takes a similar approach.<\/p>\n \u201cI have a project in Claude with a bunch of my most representative posts, as well as a voice guide I created from them,\u201d says Collins. \u201cIn fact, I think you can feed posts into Claude and ask it to create a detailed and specific guide anyone could use, and it will do a pretty good job.\u201d<\/p>\n Pro tip: <\/strong>When gathering examples, don\u2019t just dump them in. Instead, try the following.<\/p>\n Once AI has reviewed your examples, it\u2019s time to review its<\/em> understanding of your voice, not just the copy it produces. At this stage, you\u2019re looking at how well it can describe your tone, style, structure, and perspective.<\/p>\n I\u2019d encourage you to think about this step like reviewing a junior copywriter\u2019s notes before they ever write a draft to be sure they truly understand your voice before you hand them a project.<\/p>\n When I was working on the client messaging project, the AI’s first voice analysis was surprisingly strong. It nailed the overall tone, structure, and product focus. But, it missed one critical piece: the customer-first ethos that runs through every communication.<\/p>\n That might sound small, but in an industry that\u2019s not exactly known for customer obsession, it was the most important differentiator<\/em> we needed to highlight. By providing this feedback to ChatGPT and offering some qualifying questions to use as a lens, it took the comms guidelines from okay to great.<\/p>\n The importance of strong setup and thoughtful prompting came up again when I spoke with Ana Mendes<\/strong>, marketing coordinator at MeetEdgar<\/a>.<\/p>\n \u201cOur biggest takeaway? AI can be incredibly helpful, especially for small or busy teams, but only when it\u2019s trained properly and paired with thoughtful prompts,\u201d Mendes shares.<\/p>\n Pro tip: <\/strong>When reviewing AI\u2019s voice analysis, focus not just on what sounds<\/em> good but on what truly matters to your brand. The better you articulate what\u2019s missing or what needs emphasis, the faster AI will start thinking the way you do.<\/p>\n Once you’ve reviewed and refined AI\u2019s understanding of your voice, it\u2019s time to get it on paper \u2014 or at least into your AI tool.<\/p>\n For this step, I want you to take the analysis of your examples, use them to document the tone, structure, themes, and key principles that define how you write and think, and<\/em> the common traps to avoid.<\/p>\n This doesn\u2019t need to be long or fancy; it’s just clear enough that if someone new (or your AI) picked it up, they could immediately \u201chear\u201d you.<\/p>\n For example, after refining the voice analysis for my client with multiple sub-brands, I built a simple set of guidelines that includes the following.<\/p>\n There\u2019s, of course, a great deal more information that I made sure to include, but this should give you an idea of where to start. From there, you can always develop some \u201csay this, not that\u201d options as well.<\/p>\n Collins explains that documenting his voice for AI changed everything: \u201cI created a detailed, specific voice guide based on my strongest posts \u2014 and it made a huge difference in how well AI could recreate my style consistently.\u201d<\/p>\n Pro tip: <\/strong>Save your guidelines somewhere you (and AI) can easily reference them every time you kick off a new project. I\u2019ve found AI to be forgetful, so it\u2019s a good idea to keep these handy.<\/p>\n Training your AI isn\u2019t a one-and-done project. I think the junior copywriter analogy is really strong here because it requires constant tweaking and training to get your AI to write more like you.<\/p>\n That means, once you build your guidelines, start using them. Prompt AI using your framework, review the outputs, and keep offering micro-feedback as you go.<\/p>\n You\u2019ll probably notice that the more you work with your AI, the better it gets at anticipating your needs and catching your voice.<\/p>\n But you\u2019ll also notice moments where you still need to tweak, tighten, or reframe the guidelines to keep evolving as you do. In fact, I don\u2019t think this will ever change \u2014 no matter how well you train the AI, there are plenty of times that it still \u2026 sounds like AI.<\/em><\/p>\n My friend Lindsay Hope<\/a> is an AI and email strategist, and she puts this so well.<\/p>\n \u201cAI copy performs best when it\u2019s tweaked with the human touch. You still have to tie any writing back to a human, and that\u2019s what I see people miss all the time,\u201d she says.<\/p>\n The vast majority of people I spoke with agreed that AI rarely gets it right from the get-go and still needs a human editor.<\/p>\n Pro tip: <\/strong>One of the best pieces of advice I found in this came from Phillip DeRenzo, head of marketing at iAsk.ai<\/a>. He writes, \u201cRead your output out loud. If it makes you cringe, rework it.\u201d<\/p>\n When you\u2019re first training AI to capture your brand voice, it\u2019s easy to get stuck thinking you have to have a perfect voice guide ready. That might have been true at one point, but I no longer believe it\u2019s necessary, at least as you get started. Instead, I\u2019d recommend setting the stage and refining from there.<\/p>\n These prompts should work across most major AI platforms, including ChatGPT, Claude, and HubSpot\u2019s content assistant. No matter which tool you prefer, you can follow this process to start training your AI more effectively.<\/p>\n Start by giving AI a short, clear intro to the brand and audience:<\/p>\n Note: If you have done some brand voice work either on your own or with an expert like Blackman, then feel free to share more information, but don\u2019t let not<\/em> having this stop you.<\/p>\n Rather than just dumping samples in, help AI understand why<\/em> you\u2019re sharing each piece. Here\u2019s what that might look like:<\/p>\n Before rushing into analysis, encourage AI to ask if it has any questions about the goals or requested outcome (and then answer them):<\/p>\n Now, this is where I ask AI to identify patterns \u2014 and if applicable, pull out sample structures<\/em> it sees across different content types.<\/p>\n Here\u2019s what ChatGPT came up with after I asked it to analyze some of my content:<\/p>\n This is where you get to sharpen your copy chief skills and treat AI like a junior writer you\u2019re coaching. What did AI get right? What did it miss? What needs tweaking?<\/p>\n Finally, once the analysis feels solid, ask AI to turn it into a usable set of guidelines that it or a junior copywriter could use to start writing for this brand. Be sure to let AI know to name this as \u201cYour Brand Name Voice Guidelines\u201d so that you can ask for a recall anytime you start a project.<\/p>\n Here\u2019s a screenshot of the sample guide it developed:<\/p>\n Pro tip:<\/strong> If you\u2019re using HubSpot\u2019s AI Content Assistant<\/a>, their brand voice features can help reinforce consistency across blog posts, emails, landing pages, and more. It\u2019s another great way to layer in your voice as you scale content creation.<\/p>\n Now, you\u2019re ready to go! Here\u2019s how I prompt when I\u2019m ready to pick up a new project:<\/p>\n Pro tip: <\/strong>Create a dedicated project for each brand voice you train. In ChatGPT, you can also create custom GPTs to use, where you save your examples, guidelines, updates, and clarifications all in one place for easy access later.<\/p>\n Here are some examples of LinkedIn hooks GPT created using that voice (which, ahem, is mine):<\/p>\n I asked it to proceed with the first option, and this is what it cranked out.<\/p>\n I loved it. You can expect to see it soon on my LinkedIn<\/a> (or something VERY close).<\/p>\n I\u2019ve found that the examples you share matter just as much as the prompts you use.<\/p>\n Remember, AI can only learn what you actually give it. So if you want it to pick up your best tone, structure, and emotional resonance, you have to feed it strong examples from the start.<\/p>\n Here\u2019s what I typically upload (and what you should consider too):<\/p>\n Pro tip<\/strong>: Don\u2019t overload the AI with junk. Picking five to ten great<\/em> pieces that show your best work will help you get your best results.<\/p>\n <\/a> <\/p>\n AI can do a lot. But at the end of the day, it\u2019s still a tool, not a mind reader, a brand strategist, and it\u2019s definitely not a substitute for your lived experience or personal perspective.<\/p>\n Even when you train AI really<\/em> well, there are still limitations you have to work with (and around). Trust me \u2014 been there, done that, got the wrong t-shirt.<\/p>\n Here\u2019s what I recommend watching for:<\/p>\n HubSpot\u2019s Amy Marino <\/strong>emphasizes the importance of treating AI collaboration as an ongoing, iterative process. She notes, \u201cEmbrace iterative collaboration: Be specific about what\u2018s working and what\u2019s not. Direct feedback like \u2018X doesn’t make sense because Y\u2019 or \u2018too run-on, not sharp enough\u2019 helped AI course-correct quickly.\u201d<\/p>\n Marino notes that by specifying constraints (like a 30-second time limit or no more than 10 words for a headline), you\u2019ll get more practical suggestions that will fit your asset requirements.<\/p>\n As Justin Blackman<\/strong> points out, “The biggest limitation of AI copywriting is that even if you can get it to write like you, it can\u2019t think like you. So it\u2019s on you to make sure it knows your perspective.\u201d<\/p>\n Blackman notes that AI can come up with ideas, but these ideas may not be aligned with your thinking unless you give it that info. Then you can bring it to life in a way that’s uniquely you.<\/p>\n In other words, AI can learn how you sound. But it can\u2019t decide what you believe. You still have to teach it your perspective, your values, and your take on what matters most.<\/p>\n Hope<\/strong>, who\u2019s been integrating AI tools into her workflow for years, made this point crystal clear. Hope says, \u201cAI copy performs best when it\u2019s tweaked with the human touch. Anything purely informational and lacking story just feels robotic, and audiences can tell.\u201d<\/p>\n AI can save you time by giving you a strong starting point. But it\u2019s the personal anecdotes, the humor, the emotional nuance \u2014 the things only you<\/em> can add \u2014 that turn decent drafts into solid copy.<\/p>\n<\/a><\/p>\n
\n
\n
\n
Picking the Right AI Platform<\/strong><\/h2>\n
\n
Training an AI Platform<\/strong><\/h2>\n
Putting It Into Action<\/h3>\n
Steps to Train an AI Platform to Write Like You<\/h3>\n
Step 1: Tell AI who you are and what matters to you.<\/strong><\/h4>\n
<\/p>\n
\n
Step 2: Prompt AI to analyze your examples.<\/strong><\/h4>\n
<\/p>\n
\n
Step 3: Offer feedback and reiterate.<\/strong><\/h4>\n
\n
Step 4: Develop guidelines.<\/strong><\/h4>\n
\n
<\/p>\n
Step 5: Put it into action and iterate.<\/strong><\/h4>\n
<\/p>\n
Prompts to Use<\/h3>\n
Prompt 1: Introduce the brand and context.<\/strong><\/h4>\n
\n
<\/p>\n
Prompt 2: Upload the examples and explain the context.<\/strong><\/h4>\n
\n
\n
Prompt 3: Ask for clarifying questions before analysis.<\/strong><\/h4>\n
\n
Prompt 4: Ask for initial analysis (and sample structures if relevant).<\/strong><\/h4>\n
\n
<\/p>\n
Prompt 5: Offer feedback on the analysis.<\/strong><\/h4>\n
\n
Prompt 6: Turn it into a guide.<\/strong><\/h4>\n
\n
<\/p>\n
Prompt 7: Recall and use the guide.<\/strong><\/h4>\n
\n
<\/p>\n
<\/p>\n
Documents to Share<\/h3>\n
1. Website Pages<\/strong><\/h4>\n
\n
2. Emails<\/strong><\/h4>\n
\n
3. Social Media Posts<\/strong><\/h4>\n
\n
\n
4. Long-Form Writing Examples<\/strong><\/h4>\n
\n
5. Brand Documents (If You Have Them)<\/strong><\/h4>\n
\n
6. FAQs or Internal Training Docs<\/strong><\/h4>\n
\n
Limitations of AI Copywriting<\/strong><\/h2>\n
AI needs specific feedback.<\/h3>\n
AI can\u2019t think like you.<\/h3>\n
AI first drafts are just that \u2014 drafts.<\/h3>\n
AI needs you to lead the rhythm, not chase it.<\/h3>\n