How to Use AI in Your Business
(Without the Hype or the Mistakes)
Key Takeaways
Search engines find existing information.
AI tools generate answers based on patterns they’ve learned.
That makes AI fast and useful — but also means it can sound confident while being wrong.
Used properly, AI helps thinking. Used carelessly, it produces convincing rubbish.
AI is everywhere. It's in your inbox, on your phone, in every app, tool, and even whispering advice while you’re trying to run your business.
Used well, it saves time and sharpens thinking.
Used badly… well, let’s not go there.
Most people aren’t struggling with AI because the technology is bad.
They struggle because they misunderstand what AI actually is, and what it absolutely is not.
This is a simple guide on how to use AI properly... and how not to.
1. What AI Actually Does (And What It Doesn’t)
I’ll start with a mistake I made myself.
At first, I thought of AI as a kind of brain.
It isn’t.
AI does not:
Think
Understand
Believe
Know what’s true
Care whether something is right or wrong
What it does is much simpler.
AI:
Looks at huge amounts of human-created information
Spots patterns in how humans write and respond
Predicts the most likely next words based on your input
That’s it.
Think of AI as an extremely advanced auto-complete.
It predicts what sounds right, not what is necessarily right.
That’s why:
If you ask a vague question, you get a vague answer
If you ask a biased question, you get a biased answer
If you confidently say something wrong, it will often agree — confidently
This is why AI can sound convincing while being completely incorrect.
2. AI Is Not a Search Engine (This Is Where Most Confusion Comes From)
Looks through the internet
Finds existing web pages
Shows you links
Nothing new is created, you’re being pointed to information that already exists.
AI works differently.
When you ask AI a question, it doesn’t automatically “look it up”.
It generates a response, word by word, based on patterns it learned during training.
That response is designed to be:
Helpful
Clear
Plausible
Not verified. Not fact-checked against live sources.
This explains why AI can:
Explain things clearly
Answer quickly
Sound authoritative
And also why it can:
Be outdated
Miss nuance
Get things wrong without realising
For example
If a plumber asks AI for “the best boiler for a Victorian house in Kent”, it may give a confident answer, but that doesn’t mean it accounts for local regulations, current supplier availability, or real-world installation constraints.
Once you understand this difference, the rest of this guide clicks into place.
3. The Golden Rule: Quality In = Quality Out
Imagine asking a colleague to “sort this out” without telling them:
Who it’s for
What the goal is
What success looks like
What they must avoid
You wouldn’t expect miracles.
My dad used to say: "you get out of life what you put in." AI works exactly the same way.
Poor input → poor output
Vague input → vague output
Biased input → biased output
If you don’t like what AI gives you, don’t blame the tool.
Fix the instruction.
Better questions don’t just get better answers, they change the direction entirely.
4. AI Is Trained on Humans (So It Inherits Human Problems)
AI learns from human-made material.
And humans, last time I checked, aren’t flawless.
That means some of what AI has seen:
Outdated information, personal bias, confident guesses, and flat-out mistakes.
AI doesn’t know which is which.
So, here’s the rule that matters:
Never treat AI output as fact by default.
For example, if you’re an accountant using AI to explain tax changes, or a restaurant owner using it to write allergy information - you must verify it.
AI is excellent at drafting.
Humans are still responsible for truth.
5. Prompts Aren’t Questions — They’re Instructions
Most people talk to AI like it’s a magic answer box.
That’s the mistake.
A prompt is closer to a brief than a question. It tells AI:
What role to take
What information to use
What the goal is
What “good” looks like
A weak prompt produces generic filler.
A strong prompt turns AI into a genuinely useful assistant.
If you don’t tell AI what you want, it will guess, and guessing is where things go wrong.
6. A Simple Framework That Works: AIM
If you remember one thing, remember this:
AIM = Actor, Input, Mission
Actor | Input | Mission |
|---|---|---|
A journalist? | Notes? | Write, rewrite, summarise, explain, critique? |
A solicitor? | Drafts? | For who? |
A marketing consultant? | Facts you trust? | In what tone? |
A customer service assistant? | Examples you like? | In what format? |
This sets tone and priorities. | No input = generic output. | Make this better” isn’t a mission. |
“Rewrite this in plain English for a homeowner comparing loft conversion quotes” is.
7. Context Is Everything (And Why AI Feels Inconsistent)
AI doesn’t remember things the way humans do.
If you don’t give context, it fills the gaps with assumptions, often incorrectly.
That’swhy the same request can produce different results on different days.
The fix is deliberate context:
Who the audience is
What tone to use
What matters
What must not change
Without context, AI wanders.
With context, it behaves.
8. Bias Usually Starts with the Question
AI doesn’t invent most bias.
We do.
If you ask:
“Why is social media ruining society?”
You’ve already decided the answer.
If you want balanced output, ask balanced questions:
“What are the benefits and drawbacks…”
“What arguments exist on both sides…”
“What assumptions might be influencing this?”
Neutral questions don’t weaken answers. They make them better.
9. The Blind Taste Test Rule (Yes, Like the Pepsi Challenge)
In the 1970s, Pepsi ran the famous Pepsi Challenge.
People tasted two drinks blind — without knowing which was which. The point wasn’t to prove Pepsi was “better”.
It was to remove brand bias. That’s the lesson for AI.
Treat AI output the same way:
Don’t ask “Did AI write this?”
Ask “Is this clear?”
“Is this accurate?”
“Would I still use this if a human had written it?”
If it doesn’t stand up on its own merits, it doesn’t pass.
10. AI Sounds Confident — Even When It’s Wrong
This is one of the most dangerous things about AI..
It can:
Mix up facts
Invent statistics
Confidently cite sources that don’t exist
Contradict itself
All in a calm, authoritative tone.
Confidence is not accuracy.
If something feels off:
Ask follow-up questions
Ask what assumptions were made
Ask what needs checking
Break the task into smaller steps
If AI cites a farm in Canada as the best source for crops to grow in a garden in Kent — pause.
![[object Object] [object Object] [object Object] [object Object] [object Object]](/_next/image?url=%2F_next%2Fstatic%2Fmedia%2Fimg10.972282c1.png&w=3840&q=75&dpl=dpl_9UtBgnVQ2WDQ7dzms4rgxjUQDx8D)
A Quick Word on " Please " and
" Thank You "
British people love politeness.
We apologise to doors.
We thank cash machines.
Does AI need politeness?
No.
But here’s the interesting bit:
People who say “please” tend to slow down and phrase better instructions.
They’re clearer. More specific. Less rushed.
So politeness doesn’t help the AI.
It helps you.
Also, if the robot uprising ever happens, it can’t hurt.
12. AI Helps Thinking — It Doesn’t Replace It
AI cannot:
Judge truth
Understand full context
Take responsibility
Make ethical decisions
That part is still human.
The best results come from a simple loop:
AI drafts
Humans think
AI refines
Humans decide
Used properly, AI is a powerful tool. Used lazily, it’s a very confident mess.
A Simple Checklist Before You Press Enter
Ask yourself:
Is my instruction clear?
Is it neutral?
Have I defined the task?
Have I defined the audience?
Have I given enough context?
Am I prepared to check the result?
If yes, you’re using AI correctly.
13. How We Actually Use AI at Make Me Local
At Make Me Local, we use AI as a tool, not a shortcut, and definitely not a replacement for thinking.
We use it to speed up early-stage work: outlining content, organising ideas, summarising research, and helping turn technical thinking into plain English. It helps us move faster, but it doesn’t make decisions for us.
For client work, AI never runs unchecked. Everything is reviewed by a human who understands the client, their industry, and what actually matters to their business, whether that’s a plumber needing more booked jobs, a restaurant needing more covers, or a trades business trying to grow without wasting money.
AI helps us think better and work smarter. Humans still check the facts, apply judgement, and take responsibility for the outcome.
That’s the line we don’t cross.
Final Thought
AI isn’t magic.
It isn’t truth.
It isn’t judgement.
It’s a tool.
Used carelessly, it creates noise.
Used thoughtfully, it saves time and sharpens ideas.
The difference isn’t the technology.
It’s the human on the keyboard.
If you want to learn more about using AI responsibly for your business, check out our Knowledge Hub for guides, tips, and our Jargon Decoder Handbook.
Or, if you’d like some personalised advice on how AI can support your marketing without taking over, book a free marketing strategy session
Nathan Kelsey
Founder & CEO, Make Me Local
Nathan Kelsey is the founder and CEO of Make Me Local, a digital marketing agency helping service-based businesses grow through clarity-led strategies across AI SEO, web design, PPC, and content marketing. Since founding the agency in 2013, Nathan has grown Make Me Local into a global team serving over 300 businesses. Based in West Wickham, Kent, he is passionate about cutting through digital marketing jargon and building long-term growth partnerships with clients.