How to Use ChatGPT for Your Small Business Without Publishing Garbage

Most small business owners using ChatGPT are treating it like a vending machine. Put in a vague request, get back something generic, publish it, wonder why it's not working.

I've seen it happen in Holland, in Zeeland, in Grand Rapids -- good businesses with real things to say, getting mediocre output because nobody told them this part: the quality of what you get out is almost entirely determined by the quality of what you put in.

This post is about closing that gap.

Why You're Getting Generic Results

The problem isn't ChatGPT. The problem is that you're asking it to make decisions that are yours to make.

When you type "write a Facebook post about my HVAC company," you've given it almost no information. It doesn't know where you operate, who your ideal customer is, what season it is, whether you want people to call or click, what makes you different from the four other HVAC companies in town, or what tone feels right for your brand.

So it guesses. And its guess is the most statistically average version of what an HVAC Facebook post looks like. Which is exactly what every other HVAC company using AI is also publishing.

That's the vending machine problem. Vague input, average output, indistinguishable from the competition.

The fix is specificity. You have to bring the context that AI doesn't have.

How to Write a Prompt That Actually Works

Here's a before and after to make this concrete.

Vague prompt: "Write a blog post about fall lawn care."

Better prompt: "I run a family-owned lawn care company called Lakeshore Green in Holland, Michigan. I've been in business for 14 years. My customers are mostly homeowners in the $300K to $600K range who care about curb appeal and want a company they can trust, not the lowest price. Write a blog post explaining the five things homeowners should do for their lawn before winter sets in. Keep the tone friendly and practical, not technical. Avoid jargon. No bullet point overload. I want it to read like advice from a neighbor who knows what they're talking about."

The second prompt gets you something you can actually use, or at least edit into something useful. The first prompt gets you a listicle that could have been published by anyone.

A few things to always include in your prompts:

Your business context. Name, location, how long you've been around, who you serve. Don't assume it knows anything about you.

Your audience. Who is reading this? What do they care about? What are they worried about? What decisions are they trying to make?

The purpose of the piece. Are you trying to get phone calls? Build trust? Answer a question people search for? Explain a service? The purpose shapes the whole piece.

Tone instructions. Tell it specifically what you want. "Conversational, plain language" means something. "Friendly but professional" is vague enough that AI will default to boring. If you have existing content you like, paste an example and say "write in a tone similar to this."

What to avoid. I tell it to skip corporate buzzwords every single time. I tell it not to use dramatic one-sentence paragraphs. I tell it I don't want a summary paragraph at the end that restates everything. Be specific about what you don't want.

The Edit Pass Is Not Optional

Here's the part most guides skip: you have to edit everything. Not a light proofread. A real edit.

You're looking for a few specific things.

Does it sound like you? Read it out loud. If you wouldn't say it that way in a conversation with a customer, it needs to change. AI has a tendency to add phrases like "it's important to note" and "as we can see" and "a trusted partner in your success." None of that sounds like a human who runs a real business. Cut it.

Is it accurate? AI will occasionally state things that are wrong, especially anything specific: stats, regulations, processes, prices. If it mentions that something costs "an average of X" or that a law requires Y, verify that before you publish it. Not every post has factual claims that could cause problems, but the ones that do need to be checked.

Is it specific enough? AI produces content that works for any similar business in any location. Your customers are reading it to find out about you specifically. Add the local details, the years of experience, the real reasons customers choose you. These specifics are what make the content worth reading and what Google increasingly rewards.

Does it have a point? Sometimes AI produces content that is technically about the topic but doesn't actually say anything useful. It covers the ground without giving the reader anything actionable or interesting. If you read it and think "okay, so what," it needs more substance. Add your actual opinion, a real example, something concrete.

What to Use ChatGPT For vs. What to Do Yourself

Not every marketing task benefits equally from AI assistance. Here's how I think about it.

Use AI for:

First drafts of anything where you know what you want to say but getting it written is the slow part. Blog posts, email newsletters, service page copy, social captions. Get the draft, edit it into something good.

Subject line testing. Generate 10 to 15 variations on an email subject line, pick the two or three that feel right, test them. The volume AI can produce makes this actually useful.

Repurposing. Take something you've already written or said and ask AI to rework it for a different format. Turn a blog post into a short email. Pull social posts out of a webinar transcript. This kind of formatting work used to eat time; now it's fast.

Brainstorming. You're stuck on what to write about, or you need angles for a campaign, or you want to think through how to position a new service. AI is a decent thinking partner for this kind of open-ended work.

Do yourself:

Anything that requires your specific opinion or experience. If the value of the content is "what does someone with 15 years in this industry actually think," that has to come from you. AI can frame the piece, but the substance has to be yours.

Anything requiring local knowledge. Specific neighborhoods, specific competitors, specific market dynamics, recent local news that's relevant. AI doesn't have this and its guesses will be generic at best.

Customer stories and case studies. These are yours to tell, and they're some of your most valuable content. The details that make them real -- the specific situation, what went wrong, what you did, what the customer said afterward -- those can't be fabricated. Write them yourself or have a real conversation recorded and transcribed.

Anything going out under your personal name with your personal credibility attached. If it's a personal LinkedIn post or a note to your customer list, it should sound like you wrote it. That doesn't mean you can't use AI as a starting point, but the final product should be something you'd genuinely say.

The Setup That Makes This Sustainable

If you're going to use ChatGPT regularly for marketing, it's worth investing 30 minutes to set it up properly rather than re-explaining your business in every session.

In ChatGPT, you can set custom instructions that persist across conversations (look in Settings). Write a paragraph or two about your business: who you are, what you do, where you operate, who your customers are, and how you like to communicate. This becomes the default context for every prompt, which means you don't have to include it every time.

Keep a short document with a few examples of content you've been happy with. When you're working on something new and want it to match that style, paste an example in with your prompt. This trains the output faster than any written description.

And keep a list of the specific things you always have to edit out, so you can put them in every prompt. If AI consistently adds something you always delete, just tell it not to at the start. Takes five seconds and saves you the cleanup every time.

The Bottom Line

ChatGPT for small business marketing is genuinely useful. It's not magic, and the businesses treating it like magic are publishing content that nobody wants to read.

The businesses getting real value are the ones doing the work that AI can't do: providing context, editing for voice and accuracy, adding the specific details that make content feel real, and making the judgment calls about what their customers actually need to hear.

AI does the typing. You do the thinking. That's the combination that actually works.