Most people use ChatGPT or Microsoft Copilot (or Gemini, or Claude) the same way they use Google —
type a few words, hit Enter, hope for the best. And then wonder why the answer
feels generic, off-topic, or just not quite right.
Here’s the thing: it’s almost never the AI’s fault. It’s the question.
The difference between a useful AI response and a frustrating one usually
comes down to how much you tell it. Google is designed to guess what you mean
from a few keywords. AI tools like ChatGPT are designed to follow instructions
— and the better your instructions, the better your result.
The Problem With Treating AI Like a Search Engine
When you type “email template” into Google, it works fine. You get a list of
pages with email templates. You pick one, copy it, done.
Type “email template” into ChatGPT and you get… an email template. A bland,
generic one that doesn’t sound like you, doesn’t fit your situation, and
you’ll end up rewriting half of it anyway.
That’s not the AI failing. That’s you asking Google-style when you have a much
more powerful tool in front of you.
The 3 Ingredients of a Good Prompt
You don’t need to become a “prompt engineer” or learn anything technical. You
just need to give the AI three things:
- Context — what’s the situation? Tell it who you are, who you’re talking to, or what’s going on. You don’t need an essay, just a sentence.
▎ “I’m a project manager at a small logistics company…” - The task — what do you actually want? Be specific.
Not:
▎ “write an email”
but:
▎ “write a follow-up email to a client.” - Constraints — any limits or preferences? Tone, length, format. “Keep it under 100 words” or “make it friendly but professional” goes a long way. Put it together and instead of:
▎ “email template”
Try:
▎ “Write a short, friendly follow-up email to a client who hasn’t responded to my proposal in two weeks. Don’t be pushy — just check in and offer to answer any questions. Keep it under 100 words.”
That second prompt takes 20 extra seconds to write. The output is something
you could actually send.
And don’t expect AI to be perfect. Add your finishing touches. That is the satisfying part, you can focus on making the email your own after AI has done the heavy lifting.
3 Examples From a Typical Office Day
Scenario 1: You need to summarise a long email thread before a meeting
Instead of:
▎ “summarise this email”
Try:
▎ “Summarise this email thread in 3 bullet points. I need the key decisions
made, any outstanding actions, and who is responsible for each. I’m preparing
for a 15-minute catch-up meeting.”
Scenario 2: You need to push back on a request without causing offence
Instead of:
▎ “help me write a professional email”
Try:
▎ “Help me write a polite but firm email declining a request to take on
extra work this month. I want to sound helpful and offer an alternative
timeline without sounding like I’m just saying no. Keep the tone warm.”
Scenario 3: You’re preparing talking points for a presentation
Instead of:
▎ “presentation tips”
Try:
▎ “I’m presenting our Q1 sales results to senior management next week. The
numbers are mixed — some targets hit, some missed. Give me 5 talking points
that acknowledge the shortfalls honestly but keep the focus on what we’re
doing differently in Q2.”
In each case, the extra detail isn’t busywork — it’s what lets the AI actually
help you.
One Habit to Start Today
Before you hit Enter on your next AI prompt, ask yourself one question:
▎ “Would a colleague know exactly what I need from this?
If the answer is no, add one more sentence. That’s it. You don’t need a
perfect prompt — you just need enough context that a reasonably smart person
could help you.
AI tools are at their best when you treat them less like a search engine and
more like a capable colleague you can brief properly. The more you tell them,
the more useful they become.
Next Post: how to get better results by asking the AI to think step by step —
and why it makes a surprising difference.
Leave a Reply