AI can give you some of the ingredients, but it can't give you the recipe. It'll draft the plan and the follow-up email in seconds. It can't tell you whether that plan fits your business, read the politics inside your client's company, or notice you've avoided the same task for a month. Use it where you can judge the output.
Use AI for drafting and thinking out loud, and get a real person to tell you whether the plan is right.
AI answers the question you asked it. The expensive gap is the question you didn't know to ask.
Take the last strategy AI gave you and show it to someone who has signed the kind of client you want.
Even a powerful tool in the hands of a novice is just a piece of metal.
Leah Neaderthal, Smart Gets Paid
AI can do a lot for a consulting business, and I want to say that first. I'm not anti-AI. The women I work with who use it well are getting real leverage out of it.
Hand it these, without hesitation:
Here's the pattern I see: AI works best when you already know what "good" looks like. The test before you delegate anything: could you tell if it came back wrong? If yes, hand it over. If no, that's the work.
There are three things AI can't do while you're getting clients, and they decide whether you sign the work.
A B2B consulting sale runs three to 18 months, with several decision-makers, internal politics, and someone who never says their real objection out loud. AI doesn't know your client. It can't hear that the person who loves you isn't the person who signs, or that the delay isn't about you.
A client who uses AI more than anyone I know put it this way: "Leah, this is the stuff that AI could never do. Your ability to truly hear the situation with all your years of experience and know what to do, and how to coach me, not just tell me what to do, that's irreplaceable."
High-level sales judgment is oral history. It gets passed person to person, over years, because nobody can write the book that covers every situation. AI trained on what got published.
It doesn't know you talk yourself down. It won't notice you've priced this engagement below the last one, or rewritten the same email nine times because sending it feels like asking for something. It can't get you comfortable putting a number on work you used to hand over for free.
AI's advice sounds confident whether it's right or wrong, and that's the problem. There's no tell. It uses the same tone for a solid recommendation and for something it invented, and it won't stop to say "you're asking the wrong question here."
OpenAI's own researchers named the mechanism in a 2025 paper, "Why Language Models Hallucinate": models "guess when uncertain, producing plausible yet incorrect statements instead of admitting uncertainty," because the way they're trained rewards a guess over an honest "I don't know."
For most things that's survivable. Here it isn't. The trouble with the Messy Middle is that you don't yet know what you don't know, so you can't grade the answer you just got. You have no reliable way to tell good advice from confident advice.
A good coach pushes back. She'll tell you the plan is wrong, or that you're worried about the wrong thing.
Three ways to pressure-test what it gives you:
Knowing how to get clients and doing it are different problems, and only one of them is an information problem.
Most women I talk to can describe what they should be doing. Show up consistently, reach out to people who already know their work, follow up more than feels comfortable, ask for the business. The secret is there is no secret. So more information doesn't move you, and AI is information.
What moves you is someone noticing. AI remembers everything you tell it, so memory isn't the gap. The gap is that nobody says "you told me you'd send those five emails, and it's been a month."
And it's lonely. There's a specific relief in a room of women running businesses like yours, watching one of them handle the exact objection you got last week.
If you're going to keep working with AI, build the missing half:
Using AI can make your marketing worse, and I want to name the mechanism rather than be vague.
You stop listening to the instinct you do have. One woman in the Academy is as AI-forward as anyone I know. She got feedback that a section of her writing should move higher up, so she had AI do it. Her words: it was a disaster. She went back to the version she'd written by hand and thought, "oh, this is the one."
The real risk is bigger than a clunky sentence: your voice is what makes a potential client feel like they've found the person who gets it.
The order that protects it:
Real talk: a woman came to me a while back. Work had slowed. Somewhere in there she said, essentially, "ChatGPT made me a strategy and I can follow that." So she went off to run it. Three weeks later she came back and asked to work with me.
I think about that a lot. She's smart, and the strategy probably wasn't bad. But whatever she needed showed up inside three weeks, and it wasn't a missing tactic. It was not knowing whether she was doing the right thing, with nobody to ask.
Of course you'd try AI first. Nobody ever taught you to sell, and here's a tool that answers instantly and never makes you feel behind for asking.
Here's how I hold it. Even a powerful tool in the hands of a novice is just a piece of metal. Once you've got a marketing system you trust and know how to lead a sales process to a yes, AI makes you sharper at all of it. My job is to make you the samurai. Then the sword is worth having.
The Academy is where I teach both systems, with a room of women doing the same work.
No. Use it. I use it. The line I'd draw is between work you can judge and work you can't. Drafting, organizing, summarizing, prepping: hand all of it over. Deciding whether a strategy fits your business: that's judgment, and you want a person for it until it's yours.
AI can write LinkedIn posts. Whether you should depends on how well you can already hear your own voice on the page. If you can read a draft and know instantly what's yours and what isn't, AI is a fine first-draft machine. If you're still finding your voice, writing it yourself is part of finding it.
Sometimes the words overlap. The difference is what happens next. AI hands you an answer and moves on. A coach tells you the plan is wrong before you spend two months on it, and knows enough about your business to tell you which of the ten reasonable options is right for you.
AI can tell you what's typical in your category, which is worth knowing. What it can't do is look at your specific engagement, your client's budget cycle, and the value you're delivering, then tell you the number is too low. Pricing is where I see the most head trash.