How to Use AI to Create Marketing Content That Sounds Like You
Liam Sandford
Liam Sandford is a Head of Marketing, public speaking expert, and 2x Best Selling Author including the book Effortless Public Speaking. He helps ambitious professionals and entrepreneurs communicate with impact to get noticed, grow their career, and build their business.
The practical difference between using AI and abusing it comes down to method. You can use AI to create content that sounds like you, if you approach it systematically. This is the tactical guide to training AI on your voice, editing output that doesn’t match, using AI for structure while maintaining personality, and running quality control. This is distinct from the philosophy of when to use AI, this is the how.
The system works because it starts with your voice, uses AI to amplify it, and ends with your edit to restore it. When you follow these steps, AI becomes a production tool that actually serves your brand.
How to Create a Voice Training Document for AI
Start with a document of your actual writing. Pull five to ten of your best pieces. Pull pieces where your voice came through strongest, where readers responded well, and where you felt most like yourself. Include blog posts, emails, social media, or anything that represents how you think and communicate.
Extract the key characteristics. How long are your sentences? Where do you use short punches? Where do you use longer explanations? How do you open pieces? How do you close them? Do you use questions or statements? Do you use jargon or simple language? Do you reference stories or data? Are you formal or conversational?
Write these observations down in a style guide. Not a manual, a guide. It should be 500-800 words. Include:
Two to three of your actual opening sentences and why they work
Examples of how you explain difficult concepts simply
Your favourite sentence structures
Three to five direct quotes showing your tone
What you avoid (eg. jargon, acronyms etc.)
The emotional tone you aim for (confident, warm, practical, etc.)
Whether you are more formal or conversational
Share this document with AI. Tell it: “Here’s my writing style. Study these pieces and this guide. When I ask you to write, match this voice.”
Why Feeding AI Your Actual Examples Beats Generic Instructions
“Write conversationally” is vague. “Write like this” (with examples) is concrete. AI learns from patterns. Give it patterns from your actual work.
When you show AI five examples of your writing, it sees your sentence length patterns. It sees how you use questions. It sees when you break to a new paragraph. It picks up rhythm. Generic instructions can’t convey that. Patterns can.
Do this work once. Then you have a reusable training document for every piece of content you create. Every time you use AI, start with: “Here’s my voice training document. Write in this style.”
How to Structure Your AI Prompts So Output Stays True to Your Voice
The prompt is where most people lose control. A bad prompt produces generic content. A good prompt produces content close enough to edit rather than rewrite.
Structure your prompt like this:
1. Your style training (paste it or reference it)
2. The specific goal (what type of content, what topic, what outcome)
3. The outline or key points you want covered
4. The output format (how long, what structure, any specific sections)
5. Quality standards (no jargon, short paragraphs, conversational tone, etc.)
Example: “Using my voice training document attached, write a 1,200-word blog post on [topic]. Cover these points: [outline]. Use the structure: introduction, three body sections, and conclusion. Keep paragraphs to three sentences max. Sound conversational, not formal. Include a question to the reader in the introduction.”
The more detail you give, the closer to usable the output becomes. The closer to usable, the less editing you do. The less editing you do, the more you rely on AI to sound right, which defeats the purpose.
The right balance is: detailed enough that you save time, vague enough that you still have to edit substantially.
Why Editing AI Output for Voice Requires Ruthlessness
Read AI output as a first draft. The goal isn’t to polish it. The goal is to shape it into your voice.
If a paragraph sounds generic, rewrite it. If a sentence sounds like an algorithm, replace it. If AI used a word you’d never use, change it. If the pace feels off, restructure it. If a section is missing your perspective, add it.
You should be rewriting 30-40% of the content. If you’re rewriting less than that, AI is too close to your voice (which means either the content is generic or you need to push AI less). If you’re rewriting more than 80%, it’s faster to write from scratch next time.
The rewriting is where your voice comes back. Embrace it. This is your work. AI provided the structure. You provided the thinking.
Read each paragraph and ask: Would I say this? If not, rewrite it until you would.
How to Use AI for Structure While Adding Personality
AI is excellent at structure. It knows how to organize information. It can outline. It can create frameworks. Use that.
Ask AI to: outline the main points, structure an argument, create a flow from point A to point B. Then take that structure and fill it with your thinking, your examples, your perspective.
This is different from asking AI to write the content. You’re asking it to solve a structural problem. You solve the content problem.
Example: “Outline five steps to [X] in logical order.” You get a structure. Then you write each step in your voice with your examples. Or AI writes first drafts of each step, and you rewrite them with your perspective.
This method keeps AI in its strength (pattern recognition and structure) and keeps you in yours (thinking and voice).
Why Creating an AI Editing Checklist Saves Time Long term
Make a checklist of voice elements that matter to you. Every time you use AI output, run it through the checklist.
Your checklist might include:
No jargon without explanation
Sentences average 12-15 words
Questions to the reader included
Conversational tone maintained
Short paragraphs (2-4 sentences)
Active voice, not passive
Specific examples included
Call to action clear and personal
Print this checklist. Or put it as a comment in your editing document. Run every piece through it. This makes editing faster and consistent.
How to Repurpose AI Generated Content Across Formats Without Losing Voice
Once you’ve created a piece with proper voice, repurpose it across channels. The core content stays. The format changes.
A 1,500 word blog post becomes:
Five LinkedIn posts (one key point per post)
One email (longer form, more personal)
Ten social media posts (questions, quotes, examples)
Two or three video scripts (conversational, shorter)
One podcast episode outline
When you repurpose, you are translating, not recreating. Your voice comes through each format because the source material has your voice baked in. You are not asking AI to write five LinkedIn posts. You are extracting five ideas from your blog post and adapting them.
This is where AI written content becomes valuable. One strong piece, properly written in your voice, becomes five pieces across channels. That’s amplification.
Why You Should Run Quality Control Before Publishing
Before any piece goes live, run it through quality control. Read it as your ideal customer would. Does it sound like you? Does it deliver on the promise in the headline? Does it feel complete?
Ask someone you trust to read it. Tell them: “Does this sound authentic, or does it feel like AI?” Their instinct matters. If they can’t tell it’s AI, it passed.
Time also matters. Write the piece, then read it the next day. Fresh eyes catch what AI-ness remains. You will notice generic phrases you missed. You will spot where your voice dropped out.
Quality control should take 30 minutes per 1,500 word piece. If it takes longer, the piece wasn’t close enough to your voice to begin with. Next time, invest more in the prompt or the context document.
How to Know When to Write Without AI Instead
Sometimes writing directly is faster than using AI. You know the topic. You have stories. You feel inspired. Just write.
AI is a speed tool. If it doesn’t make you faster, don’t use it. If it makes the work feel like editing rather than creating, it’s not helping.
The decision is simple: Am I faster creating from scratch or editing from AI? Some days the answer changes. Stay flexible.
For more on how voice carries across all your marketing, see The Ultimate Guide to Public Speaking for Business Growth. Your content voice and your speaking voice are connected. When they are aligned, your marketing becomes exponentially more powerful. AI can help you produce more content, but your authentic voice in all formats is what drives results.
Frequently Asked Questions About Creating Content with AI
How long does it take to train AI on my voice?
The first time, 2-3 hours (extracting samples, writing the style guide, creating the training document). After that, 5 minutes per piece (pasting the training document into your prompt). The upfront work pays for itself after three to four pieces. After that, you’re saving time on every piece.
What if AI still doesn’t sound like me after following these steps?
Your training document might be too vague, or your examples might not be representative of your best voice. Pull five new pieces that feel more like you. Rewrite your style guide with more specific observations. Try again. Or the topic might require more of your voice than AI can provide. In that case, write directly and use AI for other pieces.
Should I edit before or after I finish the entire piece?
Edit as you go. Read each section as AI produces it. If it’s off, give AI feedback in the next prompt. This is faster than writing the whole piece, then rewriting half of it. One section at a time keeps quality higher and saves editing time overall.
Can I use the same context document for all content types?
You can but you won’t get as strong an output. You should have several context documents that AI can read based on the specific scenario that it needs. For example, your blog voice and your email voice should be similar but not identical. Your brand guidelines document should be the foundation, but your prompts should specify the format, skills and frameworks you are using and adjust expectations for the output.
TL;DR: How to Use AI to Create Marketing Content That Sounds Like You
Use AI as a production tool by training it on your brand voice, frameworks, and skillsets you would use if you were producing the content from scratch.
Create context documents with your actual writing examples and brand guidelines embedded
Structure detailed prompts that include your style guide, goals, outline, and quality standards
Create an editing checklist to maintain voice consistency across every piece
Run quality control before publishing. Consume the content through the lens of the viewer/reader
More From Liam Sandford
Read my book: Effortless Public Speaking. Learn how to speak confidently, reduce stress, and turn public speaking into your competitive advantage. These actionable public speaking tips will help you improve your presentation skills for any audience.
Join the free 5-day email course: Get daily lessons packed with practical strategies to deliver effective presentations and speak confidently. This course is designed to build your public speaking skills step by step. Sign up below: