You have felt it. You ask an AI tool to help with your bio, or a blog post, or a homepage, and what comes back is smooth, professional, and completely not you. It sounds like a therapist. It does not sound like this therapist. So you either spend an hour rewriting it back into your own voice, which defeats the point, or you ship the generic version and quietly wince every time you read it.
The fear that follows is reasonable and widespread. Therapists worry AI will make their content sound generic, dilute their voice, and misrepresent their expertise. They are right to worry, because with the way most people use these tools, that is exactly what happens. But the problem is not AI. It is a specific, fixable misunderstanding of how these tools work, and once you see it, you can use AI to sound more like yourself, not less.
Why does AI make therapy marketing sound generic?
AI makes your marketing sound generic because a general AI tool, given only your topic, produces the statistical average of everything ever written on that topic, and the average of all therapist writing is exactly the interchangeable middle you are trying to escape.
Understand what the tool is actually doing. When you ask a general AI to "write a therapist bio," it has no idea who you are. It has your topic, "therapist bio," and nothing else. So it generates the most statistically likely therapist bio, which is a blend of every therapist bio it has ever encountered, averaged into one smooth, professional, warm, and utterly generic result. The genericness is not a bug you can prompt your way out of with cleverer wording. It is the tool doing precisely what it is built to do: produce the center of the distribution.
And the center of the distribution is the invisible middle. It is the "same person in different sweaters" bio, generated on demand. So a therapist reaching for AI to improve their generic copy often gets copy that is more polished and no more distinct, which does nothing for the actual problem, because the tool was never given the one thing that would make the output yours: you. (This is the same sameness problem, now automated.)
Should therapists use AI for marketing at all?
Therapists should use AI for marketing, but the frame matters: used one way it flattens your voice, used another it amplifies it, and the difference is entirely about whether the tool has a model of you to work from.
Refusing AI entirely is not the answer, and it is not realistic, because the tools are becoming part of how everyone works, and a therapist who opts out entirely gives up real ground on the parts of marketing they find draining. The answer is to use AI correctly, which means understanding the one variable that determines whether it flattens you or amplifies you: what you give it to work from.
Give it only your topic, and it flattens you toward the average, because the average is all it can reach for. Give it a real model of your voice and position, and it can hold your shape, producing more of you rather than more of the middle. Same tool, opposite outcomes, and the difference is not the prompt. It is whether the tool actually knows who you are. Most therapists are using AI the first way and concluding AI is the problem, when the real issue is that they never gave it themselves to work from. (And your voice starts with a real position.)
The mistake: feeding AI facts instead of a voice
The core mistake is treating AI like a tool you inform, when you should be treating it like a tool you instruct, because facts about you do not change how it writes, but instructions about how you write do.
Here is the distinction that changes everything, and almost no one names it. There are two completely different kinds of thing you can give an AI. The first is information: facts about you, your bio, your credentials, your specialties. The AI reads this and it barely changes the output, because knowing facts about you does not tell the tool how to sound like you. The second is instruction: rules for how you actually write, think, and see, the specific things you would say and the specific things you would never say. The AI runs this, and it changes everything, because now it has a frame for how to produce, not just facts to reference.
Most therapists, trying to make AI sound like them, pile on more information. More about their background, more about their approach, more facts. And it plateaus, because information is the wrong kind of input. The output gets more detailed and no more you, because you never gave the tool the thing that governs voice: instruction about how you actually write, not information about who you are. This is why "I told it all about myself and it still sounds generic" is such a common experience. More facts was never going to work. (The full mechanism of this gets its own piece.)
How do you make AI sound like you specifically?
You make AI sound like you by giving it a real model of your voice and position to run on, the specific way you write and see, not just facts about your practice, so it has a frame to produce from instead of an average to fall back on.
The practical version, in order of impact.
First, know your actual voice and position. You cannot instruct a tool in a voice you have not identified. Before AI can sound like you, you have to know what "you" sounds like: your real position, the specific person you serve, the particular way you see them, and the concrete markers of your voice, the phrases you use, the tone you strike, the things you would never say. This is the positioning and voice work every therapist needs regardless of AI, and AI just makes having done it more valuable. (Start with the one sentence that makes you the obvious choice.)
Second, instruct the tool, do not just inform it. When you use AI, give it how-you-write rules, not just facts. "I write in short, direct sentences, I never use clinical jargon with clients, I always name the specific struggle before offering any reassurance, I would never say 'journey' or 'holistic.'" That kind of instruction is a frame the tool can run on, and it changes the output in a way that a paragraph of biographical facts never will.
Third, guard the negatives. The most overlooked instruction is what you are not. A tool told only "sound warm and professional" drifts to the generic average in every gap you did not specify, because the average is its fallback. A tool told precisely what you never say, the exact phrases, the clichés, the tones you avoid, has guardrails that hold your shape even across a long piece where it would otherwise drift back to the middle. Naming what you are not is how the voice stays yours past the first paragraph.
Fourth, always review. AI is augmentation, not replacement, especially in a field built on genuine human connection. Every piece it produces gets a human read before it goes out, because your judgment about whether it actually sounds like you and serves the client is the thing no tool has. Use it to draft and accelerate. Keep the judgment yours.
Why does my AI drift back to generic over long pieces?
Your AI drifts back to generic over long pieces because you told it what to sound like but not what to avoid, so in every gap your instruction did not cover, it falls back to the average, and a long piece is mostly gaps.
This is the failure almost no one names, and it is the felt experience of trying to use AI for anything longer than a paragraph. The opening sounds like you, because you steered it hard at the start. Then, three paragraphs in, it has quietly reverted to the smooth generic voice, because your instruction covered the general direction and not the thousand specific micro-choices a long piece requires, and in each uncovered choice it did the only thing it can: reach for the average.
The fix is negative specification, telling the tool precisely what you are not, the exact phrases and tones and moves you would never make, so that even in the gaps your positive instruction did not reach, it has guardrails instead of a slide back to the middle. This is the difference between getting your voice right for one sentence and holding it across a whole article, and it is exactly the thing a general tool, given only your topic, cannot do for you. (This is the mechanism, in full.)
The stakes are higher for therapists than for anyone
For a therapist, generic AI content does not merely underperform. It directly contradicts your entire value proposition, because your whole offer is genuine human understanding, and generic content signals its absence.
Sit with why this matters more in your field than almost any other. A person choosing a therapist is choosing a specific human they can feel safe being understood by. Generic content, content that sounds like every other therapist, actively signals the opposite of what you are selling. It whispers "interchangeable" at the exact moment the client is scanning for "this specific person gets me." So the therapist who fills their site with generic AI content is not just failing to stand out. They are broadcasting a signal that contradicts their core offer, telling the very person seeking genuine understanding that there may not be a specific human here at all. The medium becomes the message, and the message is the wrong one. (This is the structural reason a practice goes invisible.)
Where BNA72 fits
Everything above is the how. If you want the built version, the model of your voice that lets AI hold your shape instead of flattening you, that is what BNA72 is.
It starts with an assessment that maps how you actually think, decide, and connect, and turns that into a written system your AI tools can run on, so the output sounds like you specifically instead of like the average therapist, and holds that shape across long pieces instead of drifting. The $197 Foundation gives you the assessment and your positioning blueprint, the model of your voice and position. The $997 Complete builds it into a full brand and marketing toolkit. And for group owners, the $1,997 Platform does this across a whole roster, giving each clinician a distinct voice their tools can hold.
That is the destination. The frame above, knowing your voice, instructing rather than informing, guarding the negatives, reviewing everything, you can start applying today.
Find where AI is flattening you
Before you change how you use AI, find out where it is currently making you generic, because most therapists cannot tell how much of their AI-assisted content has quietly drifted to the middle.
The Generic Voice Checklist walks you through ten signs that AI is flattening your voice, so you can catch the drift in your own content before your clients feel it. It is the fastest way to see whether your tools are amplifying you or averaging you.
AI is not the thing making your marketing sound generic. Using AI without giving it a model of yourself is. Give the tool your actual voice to run on, guard what you are not, keep your human judgment on every piece, and AI stops being the thing that erases your voice and becomes the thing that helps you sound more like yourself, at more scale, than you ever could alone. That is the whole difference, and it was never about the tool. It was about whether the tool had you to work from.