How to Find the Point of Your Presentation

The mistake everyone makes first

You get asked to present. Within about ninety seconds you have opened your slide software, picked a template, and typed a title on slide one. It feels like progress. It is not. It is decorating a house you have not designed yet.

Almost every bad presentation you have ever sat through was bad for the same reason. The person who made it never decided what they wanted. They gathered material, they arranged it, they put it on a screen, and forty minutes later the room walked out carrying nothing.

The fix is unglamorous and takes about ten minutes. Before you touch a single slide, write down one sentence: the one thing you want this room to know or do. Everything else either supports that sentence or gets cut.

In plain English

Deck:: The whole set of slides. When someone says send me the deck, they mean the file.
The ask:: What you actually want from the audience: a decision, a budget, a signature, a change of mind.
Takeaway:: The one thing a person could repeat to a colleague the next day if you did your job.

Your one sentence

A good sentence has a verb in it and somebody doing the verb. Compare these two.

Weak: "An update on the customer support project."

Strong: "We should hire two more support staff before September, because response times are costing us renewals."

The first is a topic. The second is a point. Topics do not persuade anyone of anything. Notice that the strong version also tells you what your slides have to prove: that response times are slipping, that slipping response times lose renewals, and that two people fixes it. Your structure fell out of your sentence for free.

If you cannot write the sentence, you are not ready to build slides. That is not a failure. It usually means you need one more conversation with whoever asked you to present.

Using AI to interrogate your own point

Here is where an AI assistant earns its place. Not writing your slides. Arguing with you before you write them. It has no stake in your project, it will not soften things to spare your feelings, and it is available at eleven at night when nobody else is.

Give it your rough sentence and ask it to push back.

Prompt you can copy: pressure test your point

I have to present to [who they are and what power they have].
The point I want to make is: [your one sentence].
Ask me the five hardest questions someone in that room would ask.
Then tell me which parts of my point are assertion rather than evidence.
Do not be polite about it. I would rather hear it now than in the meeting.

Most people find two things when they run this. First, they were making three points, not one. Second, one of the three is the one that actually matters and the others are supporting detail. That is a good hour's work before you have opened any software.

Checkpoint

Write one sentence with a verb in it saying what you want this room to know or do. Slides come after, never before.

Know who is in the chairs

The same point needs different evidence in different rooms. A finance director wants the number. A team wants to know what changes for them on Monday. An investor wants to know why now and why you.

❌ Weak prompt

Prompt

Help me make a presentation about our new onboarding process.

Output

Here is a suggested outline. 1. Introduction 2. Background 3. Current process 4. New process 5. Benefits 6. Next steps 7. Questions.

A generic table of contents. It would work for any topic and any audience, which is another way of saying it works for none of them.

✅ Good prompt

Prompt

I have 15 minutes with three regional managers who think onboarding is fine and do not want more admin. I want them to agree to run the new two-day onboarding from October. Ask me what evidence would move people who are sceptical and busy, then tell me what to lead with.

Output

Sceptical, busy managers move on time saved and on peer proof. Lead with the hours their own teams currently lose to repeat questions in month one, not with the design of the new process. Show one region that already ran it. Keep the process detail in an appendix.

It knows the room, the ask, the time limit and the resistance, so the advice is about strategy rather than headings.

Prompt you can copy: read the room

Audience: [roles, seniority, what they care about, what they already believe].
Time: [15] minutes. My ask: [the decision I want].
What is likely to make this audience resist?
What single piece of evidence would move them most?
What should I not spend any time on?

Cut before you build

You almost certainly know more about your topic than you have time to say, and the temptation is to show it. Resist. Nobody has ever left a presentation wishing it had been longer.

Prompt you can copy: find the cut

Here is everything I could say in this talk:
[list it all, one item per line]
My one sentence is: [your point].
Sort these into three groups: essential to the point, useful if time allows,
and cut it. Be ruthless with the third group and explain each cut in one line.

AI will happily generate a confident, complete, well-organised presentation about a point you have not thought through. That is the trap. A tidy deck built on a vague idea is more dangerous than a messy one, because it hides the problem until somebody in the room asks a question.

You will notice we have barely mentioned slides. That is deliberate. Two lessons from now you will build them very quickly, because you will know exactly what they need to say.

📝 Quiz

Question 1 of 4

What should you do before opening your slide software?