What I got back was three paragraphs of corporate word salad that could have been about any service from any company in any industry. And my immediate reaction was: well, this is useless. AI doesn't get it.
Except. Yes. Obviously.
I had just walked up to a very sophisticated tool, given it the informational equivalent of a grunt and a gesture, and expected it to read my mind. That's not how this works. That's not how any of this works.
Here's the thing: we've spent two decades training ourselves to interact with technology by throwing keywords at it and hoping for the best. Google taught us that "pizza near me" gets results. That "best CRM for small business" is a complete query. That less is more, that context is irrelevant, that the algorithm will figure out what we mean.
And for search, that's fine. Google's job is to point you toward information that already exists. It's a librarian with very good organizational skills.
AI is not a librarian. AI is more like a very eager, very literal intern who will absolutely do whatever you ask, but who has no idea what your company does, what your voice sounds like, or what you're actually trying to accomplish.
When you treat AI like Google, you get Google-quality outputs: generic, technically accurate, and completely useless for anything that requires personality, strategy, or a pulse.
Prompting isn't searching. It's briefing.
When you sit down with a person to delegate work—a designer, a writer, a strategist—you don't just say "make a thing." You give them context. You explain what you're trying to accomplish, who the audience is, what tone you need, what you've already tried, what worked, what didn't.
You give them enough information to be useful.
That's what a good prompt does. It gives the AI enough context to generate something that doesn't sound like it was written by a committee of motivational posters.
Bad prompt:"Write a marketing email about our nonprofit's fundraising event."
Here's what that gets you: bland, generic nonprofit-speak that could be about literally any organization doing literally any event. Zero personality. Zero specificity. Just vibes and earnestness.
Better prompt: "Write a marketing email for our annual fundraising gala. We're a nonprofit that supports home repair for low-income seniors in Detroit. Our audience is past donors who care about aging with dignity and community support. The tone should be warm and direct, not overly sentimental. Include a specific story about one homeowner we helped last year, and make the call-to-action clear: buy a ticket or sponsor a table. Keep it under 200 words."
See the difference? You've given the AI a job description. You've told it who you are, who you're talking to, what you want them to do, and how you want to sound while doing it.
The worst part about treating AI like a search engine is that you're leaving all the strategy on the table. You're asking it to guess what you want instead of telling it what you need.
And here's the uncomfortable truth: if you can't articulate what you need clearly enough to prompt an AI, you probably don't have a clear strategy in the first place. The vagueness isn't a prompting problem. It's a thinking problem.
Good prompting forces clarity. It makes you name your audience, define your goal, specify your tone. It makes you decide what you're actually trying to say before you say it.
Which is, you know, what you should be doing anyway.
You don't need to become a prompt engineer. You don't need to take a course. You just need to stop treating AI like a search bar and start treating it like a collaborator who needs a decent brief.
Answer those questions in your prompt. Give the AI something to work with.
And if you can't answer those questions? Don't blame the AI. Fix your strategy first.
Next week, we're talking about brand-training your AI—because even good prompts won't save you if the AI doesn't know what your organization actually sounds like.