A few people have recently asked me how I create professional presentations fast. I thought it is worth to share it in a newsletter issue. I will go through the exact steps and will share the behind the scenes of creating a real life example: presenting a proposed hiring process to hiring managers.
A Quick Disclaimer Before We Start
I won't cover PowerPoint tips and tricks in this newsletter. This is simply how I as someone with solid experience of PowerPoint but zero design instinct, leverage AI to create presentations that look professional. And do it fast.
One more thing: there is no magic button that produces a great deck in one click. What I am sharing is a workflow. It takes some back and forth. But it is dramatically fast, and the results are consistently impressive.
The Confession
I know PowerPoint reasonably well. I can set up animations, control when elements appear, make transitions between slides, and pull off the small tricks that make a deck feel polished. But I have no eye for design. My taste in visual aesthetics is not great.
What I can do, though, and I believe all of the readers can do like me or better, is recognizing when something looks good. And with AI as a collaborator, that turns out to be enough.
The Process
My workflow has five steps. Each one builds on the last.
Step 1: Get the Content Right First
I heard once that content is king, and I believe in it. A beautifully designed presentation with weak content is still a weak presentation. So before thinking about slides, colors, or layouts, I focus entirely on what I want to say.
I start with an open conversation with Claude. For topics that have multiple angles or where I expect reasonable disagreement, I bring in GPT as well. I present GPT's response to Claude, ask for its take, push back, add my own perspective, and go back and forth until the three of us converge on a core idea. When I find myself agreeing with what both models are saying, I know I have something solid.
The output of this step is a clear point of view. For example:
What are the main options?
What are the pros and cons of each?
Which do I recommend, and why?
In this example: I am meeting with a group of hiring managers to discuss streamlining our hiring process. My goal is a short, focused deck that keeps the conversation moving, not a document people read while I talk. I opened a conversation with Claude on the Claude iOS app (voice mode, which I find faster for early brainstorming) and talked through what I wanted to cover.

My Prompt

Claude Response
Step 2: Organize Into a Content Outline
Once I know what I want to say, I work with AI, often Claude, to structure it. I decide on the number of slides and what each one covers, then ask Claude to present that as a proper outline: a title and bullet points for each slide, with sub-bullets where needed.
This is the format almost every business presentation follows: a headline, three to five main points, and occasional sub-points.
Important: at this stage, there are still no visuals. No colors, no layouts, no design decisions. Just text.
There will be back and forth here. AI rarely gets the outline right on the first try. The trick is not to get frustrated. Give clear, directional feedback and iterate. A few rounds usually gets you somewhere you are genuinely happy with.
In this example: Claude came back with a clean two-slide outline. Slide 1 covered the principles of an effective modern tech hiring process. Slide 2 laid out the proposed stages. I reviewed it, made a few adjustments, and moved on.

Slide One: Proposed Outline

Slide 2 Proposed Outline
Step 3: Apply Presentation Best Practices to the Outline
Before generating anything visual, I make sure the outline follows basic presentation rules. Regardless of whether AI or a human is building the deck:
Not text-heavy. Audiences cannot read and listen at the same time.
No excessive animation. Visual tricks distract more than they impress.
Context before detail. Set up the "why" before the "what."
Less is more. Fewer slides, sharper message.
These are not AI-specific tips. They apply to any presentation. But it is worth a deliberate check before moving to generation, because fixing the structure after the visual is already built costs more time.
Step 4: Generate the Presentation
Now I ask Claude to build the deck. The more direction I give it upfront, the better the first output. Specific instructions matter: how numbers should be displayed, whether progress should use a bar or a pie chart, which slides need emphasis. If I am unsure, I ask Claude for its recommendation based on the content and then I choose from the options what I believe suits the best.
You can do all of this through the standard Claude chat interface and download the file. If you have access to Claude Teams or Enterprise with Cowork, Claude can edit the file directly on your laptop which saves you some time in each iteration.
Either way: the first output will need work. That is expected. Give it specific directional feedback, slide by slide if needed and iterate until the look and feel is right.
In this example: Once the outline was locked, I sent Claude the generation prompt with a specific set of design instructions.

Slide Generation Prompt
Step 5: Polish, Consistency Check, and Final Review
This is where a good deck becomes a great one. Two things matter most:
Consistency. Font, font size, color palette, and background should be uniform across all slides. The title slide can differ but everything else should match. This is especially important when mixing slides from different decks, which happens often at work. Ask Claude to harmonize the design across the whole file.
A full run-through in presentation mode. Put yourself in the audience's shoes. Does the order make sense? Is the level of detail right? Are you providing context before diving into specifics? Do not skip this step.
One thing to keep in mind: you may still need to do some manual adjustments to the final result. Things like line spacing or replacing icons are quick fixes that are sometimes easier to do by hand than to explain to Claude. Speaking of icons, I like using Flaticon for this. It has a great selection of royalty-free icons available in high resolution in both SVG and PNG formats.
In this example: Here is the final output. Two slides, professional layout. Slide 1 uses icon cards instead of bullet points. Slide 2 shows the hiring stages as labeled columns with supporting detail underneath.

Slide One

Slide Two
Two Pro Tips Worth Calling Out
The iteration mindset. AI will not get it right the first time. That is not a bug, it is the nature of the medium. The skill is learning to give clear, actionable feedback so each round gets meaningfully better. Once you get used to this rhythm, it becomes fast.
The fake animation trick. If you want bullet points to appear one at a time as you speak, do not mess with PowerPoint's animation settings. Instead, ask Claude to duplicate the slide progressively: one slide with the first bullet, one with the first two, and so on. It is simple, reliable, and keeps the audience focused on exactly what you are talking about.
The Bottom Line
The limiting factor for most business presentations is not design skill; it is clarity of thought. AI does not replace that. What it does is remove the friction between having a clear idea and producing something that looks like it came from someone with a design team.
Content first. Iterate freely. Keep it consistent. Review it as your audience will. That is it.
Have a process that works differently for you? Reply to this email or drop a comment.