I define my voice in a way AI can actually follow
Here’s the thing about brand voice: telling AI to “sound confident” is like telling someone to “cook it well.” It’s too vague to be useful. AI needs concrete instructions the same way a recipe needs actual measurements.
I start with three or four traits that describe how I communicate. Not who I am as a person, but how I actually write. For me, that’s direct, practical, warm, and slightly informal. Those words guide everything else.
Then I get specific about the mechanics. I keep most sentences under twenty words. I use contractions because that’s how people actually talk. I say “you” a lot and almost never say “one should” or “it is recommended.” I also note words I overuse on purpose, like “actually” and “here’s the thing,” because those are part of my natural rhythm.
Just as important is what I avoid. I never use corporate phrases like “leverage synergies” or “circle back.” I don’t write in passive voice if I can help it. And I always tie advice back to real problems people are trying to solve, not abstract concepts.
The difference this makes is huge. Compare “maintain confidence in your messaging” with “make one clear claim in your first two sentences, then back it up.” The second version gives AI something it can actually do.
This whole definition fits on one page. It’s not a formal brand guide. It’s just clear enough that when I feed it to AI, the output sounds like something I would actually say.
I use a reusable voice brief instead of rewriting prompts from scratch
Every time I start a new chat with AI, I paste in the same short document. I call it my voice brief. It’s maybe two hundred words that tell the AI who I am, who I’m writing for, and how I want to sound.
The brief starts with basics. I’m a solo consultant who helps small nonprofits with donor strategy. My readers are executive directors who are stretched thin and suspicious of jargon. That’s enough context for the AI to avoid writing like I run a corporate agency or talk down to my audience.
Then I add my point of view. I believe donor retention matters more than flashy campaigns. I think most advice overcomplicated fundraising. These aren’t rules about grammar. They’re the opinions that make my content mine.
I also list my non-negotiables. I won’t promise quick fixes or claim there’s one right way to do things. If I’m not sure about something, I say so instead of faking authority. These guardrails keep the AI from drifting into generic marketing speak or making claims I’d never stand behind.
At the end, I include a few formatting preferences. Short paragraphs. No buzzwords like “leverage” or “synergy.” Questions are fine, but not in every paragraph.
The whole thing fits on one page. I update it maybe twice a year when my focus shifts or I notice the AI missing something important. But mostly, I just paste it in and start working. It saves me from explaining the same things over and over, and it keeps every piece of content feeling like it came from the same person.
I let AI handle structure and clarity, not my personality
I think of AI as my editorial assistant, not my ghostwriter. There’s a difference.
When I’m creating content, I hand off the structural work. AI is excellent at organizing messy thoughts into a logical outline. It can suggest three different headline options when I’m stuck. It tightens rambling sentences and finds the gaps in my explanations where I’ve accidentally skipped a step.
These are tasks where precision matters more than personality. Nobody reads an article because the transitions are uniquely mine. They read because the ideas connect and the information is clear.
But I never let AI write my opinions. I don’t ask it to generate the story about the client who completely misunderstood my pricing structure, or the time I nearly quit my business in year two. Those moments are mine. The strong claims I make about what works and what doesn’t? I write those myself.
My usual process looks like this: I’ll dump my rough thoughts into a document, sometimes just talking points or fragments. Then I ask AI to structure it into a draft with clear sections. Once I have that scaffold, I go back through and inject everything that makes it sound like me. I add the specific examples from my work. I insert my actual stance on debatable points, not the safe middle-ground version AI tends to produce.
The sensitive stuff especially stays human. Anything where tone could be misread, or where I’m addressing a common frustration in my field. AI doesn’t know what will land wrong with my specific audience. I do.
I keep a few signature elements that make my writing unmistakable
When I ask AI to help with content, I don’t want it to sound like every other AI-written piece out there. So I’ve identified a handful of things that make my writing recognizably mine, and I make sure those show up consistently.
For me, it’s about three things. First, I always explain complex ideas using cooking analogies. It’s just how my brain works. When I’m talking about layering strategies, I’ll compare it to building flavor in a sauce. When AI drafts something for me, I go back and swap out generic metaphors for ones from the kitchen.
Second, I have specific ways I name common problems my audience faces. Instead of calling something “time management issues,” I call it “the calendar chaos spiral.” These aren’t cute catchphrases for the sake of branding. They’re just how I actually talk about these problems, and my readers have started using the same language back to me.
Third, I keep a short list of personal stories I reference regularly. There’s the story about my failed product launch in 2019. The time I accidentally sent a draft email to my entire list. Maybe five or six stories total that illustrate the lessons I teach most often.
When I brief AI on a new piece, I’ll say something like “explain this using a cooking metaphor” or “reference the 2019 launch story here.” It takes an extra minute, but it keeps everything feeling like it came from the same person.
The key is these aren’t arbitrary choices. They’re patterns that already existed in how I communicate. I just identified them and made them deliberate.
I edit with a simple voice check before anything goes out
After the AI gives me a draft, I do one quick pass before I publish anything. I’m not looking for perfection. I’m looking for whether it sounds like me.
The first thing I do is read a few sentences out loud. If I stumble or cringe, that’s a sign something’s off. Would I actually say this to a client over coffee? If the answer is no, I rewrite it.
Then I scan for what I call AI-ish phrases. Things like “in today’s fast-paced world” or “it’s important to note that” or “leverage your expertise.” Nobody talks like that. When I spot them, I either cut them completely or swap them for how I’d actually explain the idea.
I also check whether the claims match what I actually believe. AI loves to be helpful and agreeable, so sometimes it’ll make my position sound more certain or more cautious than I really am. If I’m confident about something, I let that show. If I’m still figuring it out, I say that too.
The opening and closing get extra attention. That’s where bland AI writing shows up most. I tighten the first sentence so it pulls people in, and I make sure the ending feels like something I’d actually leave someone with, not just a polite summary.
This whole process takes me maybe five or ten minutes. It’s not about fixing every word. It’s about making sure the piece feels like it came from a person, not a bot trying to sound like a person.
I handle high-stakes topics differently to protect trust
I don’t treat all content the same when I’m working with AI. Some topics matter more than others, and that’s where I pump the brakes.
Anything involving pricing, guarantees, or promises gets my full attention without shortcuts. These are the sentences that could end up in a customer email or a difficult conversation later. I’ll use AI to help me say things more clearly, but I personally verify every claim and check that I’m not accidentally overpromising.
The same goes for anything that touches on legal, health, or financial topics. Even if I’m just sharing general information, I slow way down. I might ask AI to help simplify confusing language or organize my thoughts, but I’m the one deciding what level of certainty to express. If I’m not an expert on something, I say so directly rather than letting AI fill in gaps with confident-sounding statements.
Customer stories and testimonials are another careful zone. I use AI to polish the writing, but I never let it invent details or amplify what someone actually said. The trust people place in me when they share their experience is not something I’m willing to risk for faster content.
The pattern is pretty simple. AI helps me communicate more clearly, but on high-stakes topics, I’m personally verifying facts, adding context about what I do and don’t know, and making sure my tone matches the seriousness of the subject. The goal isn’t to avoid AI completely. It’s to stay in control of anything that could damage trust if I got it wrong.
