You're already paying for the tool. ChatGPT, Microsoft Copilot, Claude — one of them is sitting in a browser tab right now, probably used a handful of times with mixed results. The output was fine, maybe. Generic. Not quite what you had in mind.
That's almost never the tool's fault. It's the prompt.
A prompt is just what you type to the AI — the instructions you give it before it responds. Most people treat it like a Google search: a few words, maybe a sentence. That works fine for looking up a flight. It doesn't work for getting a draft email that sounds like you, a meeting summary with actual action items, or a vendor comparison you'd actually use.
The good news: writing a better prompt is a learnable skill, and the gap between a mediocre prompt and a genuinely useful one is smaller than you'd think. This guide will show you exactly what to put in a prompt, what to stop doing, and give you four templates you can copy and use today.
Why AI Gives You Mediocre Output (It's Not the AI)
AI language models are completion engines. They take what you give them and try to produce the most plausible continuation. The less specific your input, the more they fill the gap with averages — average tone, average structure, average assumptions about what you want.
Ask ChatGPT to "write a follow-up email to a client" and it will produce something technically correct and utterly bland, because it knows nothing about your client, your tone, the context of the conversation, or what you actually want the email to accomplish. It's the difference between generic output and a follow-up that actually sounds like you.
That's not a flaw. It's a feature working as designed. The model is doing its best with almost no information.
Your job is to give it more information.
What a Good Prompt Actually Contains
Think of a well-built prompt as having four components. You don't always need all four, but the more of them you include, the better the output.
1. Role — Tell it who to be
Start by telling the AI what perspective or expertise to bring to the task. This shapes vocabulary, tone, and the assumptions it makes.
- Weak: "Write an email to a client."
- Stronger: "You are an experienced account manager at a small B2B services firm."
You're not casting a magic spell — you're narrowing the model's frame of reference so it doesn't default to something generic.
2. Context — Tell it what's actually going on
This is the part most people skip, and it's the most important. Give the AI the background it needs: who's involved, what's happened, what the stakes are.
- Weak: "Summarize this document."
- Stronger: "I'm the owner of a 12-person HVAC company. This is a contract from a new commercial subcontractor. I need to understand the key obligations, payment terms, and anything that looks unusual or risky."
Context is what transforms generic output into something that actually fits your situation.
3. Format — Tell it what you want back
If you don't specify a format, the AI will pick one. Sometimes it guesses right. Often it doesn't. Be explicit.
- "Give me a bulleted list."
- "Write this as a short paragraph, no more than 100 words."
- "Structure this as: Subject, Key Points, Recommended Next Steps."
- "Give me three options, each one sentence."
Format instructions are free. Use them.
4. Examples or Tone — Show, don't just tell
If you want a specific voice — professional but warm, direct without being rude, formal — describe it or show it. You can even paste in a sample of your own writing and say "match this tone."
- "Write in a friendly but professional tone — the way you'd talk to a longtime client, not a stranger."
- "Here's an email I wrote last month that hit the right tone: [paste]. Write the new email in the same style."
This is the difference between output that sounds like you and output that sounds like a corporate template.
The Most Common Mistakes (and the Fix)
Mistake 1: Being too vague
"Help me with marketing" tells the AI almost nothing — least of all which marketing tasks it should actually help with. Fix: Specify the task, the audience, the deliverable, and the context.
Mistake 2: Asking for too many things at once
One prompt, one task. If you need an email and a social post and a summary of your notes, run three prompts. Stacking tasks produces mediocre everything.
Mistake 3: Accepting the first output without pushing back
AI output is a draft, not a finished product. If something's off, say so. "Make this shorter." "That's too formal — loosen it up." "The second option was closer — can you give me two more like that?" The conversation is the tool.
Mistake 4: Forgetting that the AI doesn't know your business
It doesn't know your pricing, your clients, your history, or your preferences unless you tell it. The more relevant context you provide, the less the AI has to guess.
4 Copy-Paste Templates for Real Business Tasks
These are starting points. Swap in your own details — the placeholders in brackets are yours to fill.
Template 1: Draft a Professional Email
You are a professional account manager at a small [industry] business.
I need to write an email to [client/vendor/contact description] about [situation — e.g., "a project that's running two weeks behind schedule due to a supplier delay"].
The goal of the email is to [what you want the reader to feel or do — e.g., "keep the client informed, reassure them we're on top of it, and set a new expected completion date of [date]"].
Tone: [e.g., "professional but warm — this is a client we've worked with for three years"].
Length: [e.g., "short — 4 to 6 sentences"].
Format: [e.g., "no bullet points, just a clean paragraph or two"].
Do not use filler phrases like "I hope this email finds you well."
Template 2: Summarize a Document or Meeting Notes
I'm the [your role] of a [size and type of business — e.g., "25-person manufacturing company"].
Below are [notes from a team meeting / a contract / a vendor proposal / etc.]. Please summarize it with the following structure:
1. What this is (one sentence)
2. Key points I need to know (bulleted list, plain English)
3. Any action items or deadlines
4. Anything that looks unusual, risky, or that I should ask a question about
Keep it concise. I need to be able to review this in under two minutes.
[Paste your document or notes below]
Template 3: Turn Rough Notes Into a Plan or Proposal
I have rough notes from a [meeting / brainstorm / phone call] that I need to turn into a clear [action plan / project brief / proposal outline].
Here are my notes:
[Paste your notes]
Please organize these into:
- A one-paragraph summary of the goal
- A list of key action items, each with an owner if I've mentioned one and a rough timeline if I've given one
- Any open questions or decisions that still need to be made
Write it so I could send it directly to my team. Keep the language clear and direct.
Template 4: Brainstorm Options or Approaches
I'm a [your role] at a [type and size of business]. I'm trying to figure out [the decision or problem — e.g., "how to reduce the time my team spends on client onboarding, which currently takes about 4 hours per new client"].
Give me 5 concrete approaches I could take — not general advice, but specific things a business like mine could actually try. For each one, include:
- What it involves
- Roughly how hard it is to implement (low/medium/high effort)
- The main tradeoff or risk
I'll use this list to decide what to explore further, so I don't need full plans — just enough to think clearly about the options.
One More Thing: Iterate, Don't Restart
The single best habit you can build with AI tools: treat every output as a draft in a conversation, not a final answer. When the response is close but not quite right, say exactly what's off. "The tone is too stiff." "This assumes I have a bigger team than I do." "Give me the second bullet as its own email." You'll almost always get something better in the next response than you got in the first.
Most of the owners who tell us AI "doesn't really work" for them have been abandoning conversations after the first response. That's like hiring a contractor, hearing their first idea, and firing them before you've had a real discussion.
The Payoff Is Real — But It Requires a Little Setup
Prompting well is a skill, and like most skills, the first few attempts feel clunky. Use the templates above to skip the learning curve. Once you've got a handful of prompts that work for your most common tasks, save them somewhere. Build a small library. Your prompts become assets.
If you want to go deeper — figuring out which AI tools actually make sense for your business, and how to use them in ways that save real time and money — that's the kind of thinking we do with clients. Book a strategy call and we'll take a look at what's worth your time.