Visual Thinking with AI

From mental chaos to clear action in less than 10 minutes

Today’s edition of the Augmented Advantage is co-authored by Valentin Marguet, project lead in the automotive industry.

Hi there,

Have you ever been stuck in one of those endless meetings where everyone's talking in circles about a complex problem, but nobody seems to be getting anywhere? Yeah, me too.

Recently, my co-author Valentin shared how his engineering team uses AI-powered visual thinking to break through exactly these kinds of deadlocks.

Today, we'll show you three proven examples that cut meeting time in half and drive faster decisions.

Let's dive in!

Why Visual Thinking Matters Now

Think about your last meeting. The one where you spent three hours watching your colleagues debate in circles. Everyone had their own mental model of the problem, but nobody could quite explain it. You probably left with a vague action plan, only to realize days later that half the team had a completely different understanding of what was agreed upon.

This isn't just eating up your time - it's killing your momentum.

While you're stuck in meeting purgatory, you could go ahead actually building your AI roadmap, not talking about. And the worst part is that you know if everyone could just see what you're thinking, they'd get it in seconds.

And that's where visual thinking comes in to improve:

  • Decision Making: Seeing the whole picture by mapping trade-offs, identifying hidden dependencies, and evaluating multiple scenarios simultaneously. For example, when our automotive product team faced a make-or-buy decision for a new hardware component, visualizing the decision tree revealed critical integration points that traditional analysis had missed.

  • Ideation: Breaking mental blocks by making ideas visible! This enables us to recognize patterns across seemingly unrelated concepts, iterate quickly on potential solutions, and build collaboratively on each other's thinking.

  • Communication: Going from confusion to clarity – the cost of miscommunication in business is staggering, with companies losing $12,000-$25,000 per employee annually due to poor communication. Visual thinking tackles this head-on by making complex ideas clearer, helping cross-functional teams align faster, and dramatically improving stakeholder buy-in.

And luckily, AI is great at helping us with visual thinking!

How it works

Modern AI is amazing at one thing: spotting patterns.

Whether in text, code, or abstract concepts, AI can identify underlying structures and relationships. Once it sees these patterns, it can translate them into specialized visualization languages like Mermaid or UML.

Think of it as AI turning your messy thoughts into a structured blueprint, which then gets rendered into a clear visual.

So when you explain a complex problem to ChatGPT or Claude, they don't just understand the words - they grasp the context, hierarchies, and flows. Within seconds, they can turn this understanding into visible diagrams that make sense to everyone.

This basically works in every modern LLM-based AI tool:

  • Claude AI: Creates great visuals out of the box.

  • ChatGPT: Generates code for visuals to be imported into a rendering tool.

  • Microsoft Copilot: Works, but how well depends on the version you're using.

No design skills needed!

Prompting the AI

When it comes to creating the visuals, don't overthink it. Tell the AI what you want (which type of chart) and give it some context to work with

Here's a prompt to get you started:

"Create a Mermaid flowchart showing the process of [your process]. Include decision points and key stakeholders."

Examples

Here are three (adapted) examples where visual thinking transformed complex situations:

Example 1: Quick Project Assessment

An automotive manufacturer transitioning to EV production needs to optimize their battery supply chain network across two continents, working with three major cell suppliers and managing two strategic module assembly locations.

Instead of managing complex spreadsheets, the team creates a network visualization that:

  • Maps relationships with three key cell suppliers across EU, NA, and Asia

  • Shows the streamlined battery integration process: Cell supply → Module assembly → Pack integration

  • Illustrates distribution to four major vehicle assembly plants

  • Includes realistic transport modes and times between each step

  • Highlights backup supply routes for risk mitigation

They use the following prompt in ChatGPT which describes some key parameters:

Create a Mermaid graph showing our EV battery supply chain. Include:
* Three cell suppliers (EU, NA, and Asia locations)
* Two module assembly hubs (Europe and North America)
* Pack integration centers
* Vehicle assembly plants
* Transport modes and times
* Emergency backup supply routes
Use edge labels for transport modes and times. Color code by operation type (Cell supply, Module assembly, Pack integration, Vehicle assembly).
Show both primary and backup logistics flows.

Then, they take the Mermaid code that ChatGPT generates, put it into Mermaid Live and voilà – there’s your chart:

ChatGPT also gives you the explanation of what’s happening here:

The visualization helps in a way that:

  • Stakeholders can quickly understand the overall flow from cell to vehicle

  • Supply chain risks and backup options become immediately visible

  • Transport optimization opportunities are easily identified

  • Regional capacity balancing decisions become clearer

  • Quality control points are clearly mapped across the process

All done in less than 10 minutes with AI and without any extra dollar spent.

Example 2: Crisis Response Flow

When a tier-1 automotive supplier needs to coordinate quality alerts across 3 regions, they create a sequence visualization rather than relying on complex procedures and email chains.

Instead of digging through standard operating procedures, the team visualizes their quality alert flow that:

  • Shows immediate actions for each team

  • Highlights standard and emergency paths

  • Includes realistic response times

  • Maps parallel activities to save time

They can use the following prompt directly in Claude.ai with some internal processes description:

Create a Mermaid sequence diagram for quality alerts. Include:
* Key teams: Detection, Quality, Management, Customer, Production
* Realistic time targets: 30min, 2h, 4h, 1d
* Standard and emergency flows Use dark theme and subtle highlighting for emergency protocol

The following chart was generated immediately:

The visualization helps in a way that:

  • Teams know exactly when to act

  • Response time improves significantly

  • Emergency protocols become clear

  • Communication gaps disappear

Example 3: Topic Maps

But it doesn’t always have to be technical. Visual thinking is also a great way to check the consistency and flow of ideas in any creative text. Like a sales page, a pitch document, or: this newsletter!

For example:

Create a reverse outline of this article and show the result as a mermaid flowchart which shows the logical flow of elements as well as cross-references or redundancies of topics using arrows.

[newsletter content]

And then followed by:

Simplify it. Add blue-teal shade colors for different groups. Change font to Inter. Make top-level nodes bold. Round corners. White font.

As a result, we would get this:

Looks pretty organized? Great! That’s what we want!

Getting Started Today

As mentioned before, don’t overthink this. Getting started is easy:

  1. Pick Your Tool
    Start with Claude AI – it's the most straightforward

  2. Start Small

    Take your next meeting agenda and spend 5 minutes visualizing it with this prompt:

    "Create a Mermaid diagram showing the key discussion points and their relationships for [meeting topic]"

  3. Build Your Library

    Save the prompts that work well - they're your new productivity shortcuts.

Remember: A "good enough" visualization that drives action beats a perfect PowerPoint that stays in your drafts.

Keep innovating!

See you next Friday,
Tobias & Valentin

Reply

or to participate.