ChatGPT tips

You ask ChatGPT to write a homepage. The result is polished, reasonably clear and somehow sounds like every other company in your industry. So you ask it to make the copy more professional. Then more engaging. Then less generic. Six revisions later, you’re wondering whether writing it yourself would have been quicker. Often, the missing ingredient is the brief. Learning how to use ChatGPT effectively starts with explaining what you want to achieve. A homepage for a business seeking distributor enquiries needs a different approach from one selling products directly to consumers. If you leave that distinction out, the AI has to fill the gap. That is where many disappointing results begin.

Give the task a reason

“Write me an email” is a perfectly reasonable request when you need a quick draft. But it leaves several important questions unanswered.

Who is receiving it? What has already happened? What should they do after reading it?

Consider a company whose new customers keep asking the same questions after signing up. The immediate task might be writing a welcome email. The actual problem is that customers don’t understand what happens next.

A more useful request would be:

We want new customers to complete their account setup without needing to contact support. Our current welcome email is attached, along with the five questions customers ask most often. Rewrite the email so it explains the next steps clearly and highlights anything missing from our onboarding information.

Now the writing has a purpose. It can be assessed against something more useful than whether it “sounds good”.

How to use ChatGPT effectively starts with context

You would not expect a new colleague to understand your business after hearing its name. The same applies when briefing AI.

Useful context might include your audience, the service you provide, the customer’s main concern and the limits you need to work within.

For example, a Dubai property agency could explain that it works with families relocating to the UAE, focuses on particular neighbourhoods and wants viewing enquiries rather than general website traffic.

Those details affect the topics worth covering, the questions an article should answer and the action readers should take.

OpenAI’s prompting guidance similarly recommends describing the result you need and supplying relevant context, the required output and any important boundaries. It also makes clear that every request does not need an elaborate formula.

A concise, specific brief is usually more useful than several paragraphs telling the AI to act like a world-famous expert.

Explain the whole project, then choose a sensible starting point

One reason AI-generated work can feel disconnected is that each piece begins with a different set of assumptions.

The homepage targets one audience. The sales email stresses a different benefit. The social posts use a tone nobody would recognise from the website.

Giving ChatGPT the wider project brief can help establish a consistent direction.

Imagine you are launching a business that supplies salon products to distributors. Explain the commercial model, the intended buyers, the product range and what makes the offer credible. Then ask it to propose the website structure before drafting individual pages.

You can review the structure, correct the positioning and continue with the copy.

There is no need to request an entire website, launch campaign and sales strategy in one enormous answer. Understanding the whole project and completing it in manageable stages can work together.

The aim is to avoid discovering, halfway through the job, that everything was written for the wrong customer.

Connect related tasks into a useful workflow

Once the objective is clear, it becomes easier to see how individual outputs should support one another.

Take a property agency planning content for prospective tenants. A useful sequence might begin with questions gathered from actual enquiries. Those questions could inform a content plan, which then leads to an article, a newsletter section and several social posts.

Each piece serves the same audience and addresses a recognisable concern.

That is more useful than asking for thirty unrelated post ideas simply because the calendar has thirty empty spaces.

Keep research grounded, too. If you want current competitor information, ask for web research and inspect the sources. If you want to know what your own customers ask, provide enquiry records or notes you are authorised to share. An AI-generated list of “common questions” is a starting hypothesis, not evidence about your customers.

The quality of the workflow depends on what goes into it.

A practical brief you can adapt

Here is an example for a fictional UAE cleaning company:

Help us increase enquiries for our office cleaning service over the next three months.

We serve small and medium-sized offices in Dubai. Our buyers are usually office managers who care about reliability, clear pricing and minimal disruption.

We can publish one blog and two LinkedIn posts each week. Our service information and recent customer questions are attached.

Suggest a realistic content plan and explain which customer concern each topic addresses. Then draft the first article and its two supporting posts.

Use British English and a straightforward tone. Do not invent prices, customer testimonials or performance figures. Flag any information you need from us before making a factual claim.

This gives the work a commercial purpose without pretending that content alone guarantees enquiries.

It also tells the AI what the team can realistically produce. A daily video strategy is not much use to a business with two spare hours on a Friday.

Give feedback that points to the problem

“Make it better” gives ChatGPT very little direction.

Useful feedback identifies what is wrong and what should change:

The introduction takes too long to reach the customer’s problem. Open with the issue of cleaners arriving during office hours, then explain our scheduling options.

Or:

This sounds too promotional. Keep the practical advice and remove claims we cannot support.

Specific feedback can turn a broadly acceptable draft into something that fits the business.

There is no prize for getting the first prompt perfect. What matters is recognising whether the output is moving towards your objective.

Keep judgement and execution separate

A strong brief can improve the work, but it does not make every answer correct.

You still need to check facts, review recommendations and decide whether the result reflects your business. Code needs testing. Customer-facing claims need evidence. A confident explanation of a competitor’s strategy may still be an inference.

There is also a difference between preparing work and carrying out an action. Drafting a newsletter does not mean it has been sent. Producing website copy does not mean the website is live.

What ChatGPT can execute depends on the tools, connections and permissions available. State clearly when you want a draft for review and when you want an authorised action completed.

That clarity helps prevent misunderstandings on both sides of the screen.

Start with a job that keeps coming back

You do not need to redesign your entire business around AI to get value from it.

Choose one recurring piece of work: preparing proposals, answering onboarding questions or turning technical notes into customer-friendly content.

Describe what a successful result would look like. Supply a good example, explain the constraints and review the first attempt carefully. Once the process works, reuse the brief and improve it as your needs change.

Knowing how to use ChatGPT effectively is less about collecting clever phrases and more about explaining the work clearly.

Before your next request, finish this sentence:

“By the end of this, I need…”

It is a small change, but it gives the rest of the conversation somewhere useful to go.

Looking to apply AI to everyday business processes? Talk to Snobbots about your requirements.