I have a confession: I am a little obsessed with McKinsey.
Not in a “I want to be a consultant” way. More in a “the way these people think about problems is genuinely different and I want that” kind of way.
If you have ever read a McKinsey report, sat through a McKinsey-style presentation, or studied how their analysts actually approach a market — you know what I mean. There is a discipline to it. A structure. Everything is there for a reason, nothing is filler, and the thinking is always working backward from a clear point of view.
I became obsessed with applying that to my own work in international sales years ago. And for a long time, the gap between what I wanted my presentations to look like and what I could actually produce on my own was… frustrating.
What McKinsey Thinking Actually Means (In Practice)
When most people hear “McKinsey-style,” they think of a certain visual aesthetic. Clean slides. Navy headers. Lots of white space. Minimalist charts with one clear insight per slide.
That part matters. But it is not the main thing.
The real McKinsey signature is the thinking behind the slides. The Pyramid Principle — leading with the conclusion, then building the evidence underneath it. MECE thinking — making sure your categories are Mutually Exclusive and Collectively Exhaustive, so nothing overlaps and nothing is missing. The constant discipline of asking “so what?” at every step.
In market analysis, this means you do not just collect data. You build a point of view from the data, then you structure your argument so the reader understands your conclusion before they even get to the evidence.
In presentations, this means every slide has exactly one message. The title of the slide is not a topic (“Q1 Sales Performance”) — it is a statement (“Q1 Performance Exceeded Target Despite Headwinds in Two Key Markets”). The slide proves that statement. Nothing more.
I have been trying to work this way for years. The problem was the tools never helped me get there.
Why Most AI Tools Miss This Completely
When I started using AI assistants for work, I was hoping they would help me think more rigorously, not just write faster.
That is not what I got.
ChatGPT and Gemini are genuinely good at producing content. But when I asked them to help me structure a market analysis or build a presentation, the output always had the same problem: it was organized by topic, not by argument.
I would get slide titles like “Market Overview” and “Key Challenges” and “Recommendations.” Perfectly fine categories. Completely generic. No point of view. No narrative. The kind of structure that makes a reader work hard to find the insight instead of having it handed to them clearly.
To get from that to something that actually felt like rigorous thinking, I still had to do all the hard intellectual work myself. The AI gave me a container. I still had to fill it with a real argument.
What Changed When I Started Using Claude
The first time I asked Claude to help me structure a market analysis, I described what I was working on: a regional sales review for the Middle East and Africa, with country-level data, some clear winners and some underperforming markets, and a leadership team that needed a clear recommendation on where to focus in the next quarter.
I also told it I wanted the structure to follow the Pyramid Principle. Lead with the answer. Support it with evidence. Make each section MECE.
What came back was not a topic outline. It was an argument.
The opening slide was a single statement of the recommendation. The next layer broke that recommendation into three supporting points, each one covering a distinct part of the story with no overlap. Every data point was placed where it supported a specific claim, not just dropped into a “data” section.
It was the first time I had used an AI tool and felt like the output was actually thinking, not just organizing.
The Visual Side: McKinsey Navy, Clean Layouts, No Clutter
I should talk about the aesthetic too, because it matters more than people admit.
McKinsey presentations have a very specific look. Dark navy headers. Clean white backgrounds. Charts stripped down to their essential data with no decorative elements. Generous white space. No bullet point walls. Every element on the slide has a reason to be there.
This is not just about looking professional. The visual discipline reinforces the intellectual discipline. A cluttered slide suggests cluttered thinking. A slide with one clear message and clean supporting data signals that you know exactly what you want to say.
When I describe this style to Claude — the specific color palette, the layout principles, the rule about one message per slide — it gets it immediately. Not just the description of it, but the underlying logic. It understands why the constraints exist, and that means the output actually follows them rather than paying lip service to them.
I have asked other AI tools to apply this kind of style guide. The results were always technically compliant and conceptually off. Claude produces output where the structure and the aesthetics are actually connected — where the visual choices serve the argument rather than just decorating it.
What This Looks Like When I Am Building a Real Deliverable
Here is roughly how I now work when I need to put together a serious market analysis or executive presentation:
I start by telling Claude what I am working on — the context, the audience, the key data I have, and the question leadership needs answered. Then I ask it to help me structure the argument before we build any slides. What is the headline conclusion? What are the three things that support it? What evidence belongs where?
Once the skeleton is right, I ask it to write the slide content — titles that are statements, not topics; supporting points that are specific, not generic; data callouts that drive a conclusion, not just display a number.
What I get back is something I could put in front of a room without apologizing for it. Not a rough draft that I then need to completely rethink. Something that is already doing the intellectual work the presentation is supposed to do.
That is new. That did not exist for me before Claude.
Why I Think This Works Better With Claude
My honest theory is that Claude handles nuance and context at a different level than most AI tools. When you give it a framework — not just a style preference but an actual intellectual framework like the Pyramid Principle — it does not just apply the label. It applies the logic.
That is the difference between an AI that follows instructions and one that understands them. Most tools I have tried fall into the first category. Claude, at least in my experience, is closer to the second.
For someone who cares about the quality of their thinking — not just the quality of their output — that distinction matters enormously.
The Bottom Line
If you have spent time studying how McKinsey thinks, how they structure arguments, how they approach a market — and you have wanted an AI tool that actually operates at that level — Claude is the closest thing I have found.
It will not replace the thinking you have to do. The best output still comes from a person with a clear point of view who uses the tool to sharpen and structure that thinking, not outsource it entirely.
But for the first time, I have an AI that meets me where I am — that understands why the structure matters, not just what the structure looks like.
That is rarer than it sounds.