One of the first lessons Mike Goempel taught me when I started my unexpected journey into software testing was this: Don’t just write test cases—map your mind.
Mike handed me a dry-erase marker and pulled me over to a whiteboard. “Start here,” he said, “with what you know. Then just follow the edges.”
At first, I thought this was just a quirky brainstorming exercise. But it turned out to be one of the most powerful strategies I would carry with me into client meetings, exploratory sessions, and even interviews. Mind mapping—literally drawing the web of components, inputs, flows, failure points, and user interactions—turned out to be the best way I knew to see the system.
It made things visible before they became problems.
The TestInsane Treasure Chest
Later, I discovered TestInsane’s Software Testing Mind Maps, which blew the concept wide open. Dozens of detailed maps on everything from login pages to browser compatibility to REST API testing. These weren’t just diagrams—they were distilled tester wisdom.
In moments when I felt stuck, those maps helped me ask better questions:
- What kind of data should I try here?
- What happens if the network flakes out?
- What roles haven’t I considered?
- What mental model is the user bringing to this feature?
Mind maps trained me to think like a tester—not just a checker.
And they reminded me that the real job isn’t to cover every edge case; it’s to explore the edges of what we understand about the system.
But What About AI?
Now we’re in a different moment.
There are tools that can generate hundreds of test cases from a user story. Tools that analyze your logs, learn your app flows, and even recommend tests based on statistical models. It’s easy to wonder: Do I still need to draw messy webs on a whiteboard?
My answer is yes. Maybe more than ever.
AI is good at surfacing patterns. It’s good at generating plausible test paths. But mind maps aren’t about plausibility. They’re about curiosity. They’re about turning a vague idea into a network of possibilities and risks, and noticing the areas where no arrows exist yet.
Mind Maps + Language Models: A Creative Duo
Here’s what it looks like in practice:
Let’s say I’m testing a new referral workflow in a healthcare app. I’ll pull out my iPad or a whiteboard and start a mind map with the central node: “Referral Flow”. From there I branch into:
- User roles
- Input sources
- Data dependencies
- Third-party integrations
- Notifications
- Audit trail
Now I’ve got a messy but meaningful diagram. That’s when I invite the AI in.
I might ask ChatGPT:
“Given this referral flow with X, Y, and Z components, what are some edge cases or risky transitions I should explore?”
Or:
“Can you generate test cases for the nodes I’ve outlined here?”
Even better, I can copy-paste parts of the mind map into a prompt:
“For a workflow involving user-submitted referrals, a scheduling engine, and notification logic, what are 10 test ideas involving failure states or degraded network conditions?”
What the model gives me back isn’t a replacement for the map—it’s an enhancer. A second brain to bounce against. A pattern-spotter. A list-maker. But it’s my messy mind map that gives it direction.
Without the map, AI becomes reactive—answering the wrong question really well.
With the map, AI becomes collaborative—adding depth to the shape I’m already sketching.
Mapping What Matters
These days, my maps are less polished. Sometimes they’re scribbled in GoodNotes or sketched on a Post-It. Sometimes they never leave my head. But the discipline remains: Think in branches, not in lists.
If you’re new to testing—or if AI tools are starting to make you question your instincts—I’d encourage you to try this:
Take a feature. Draw a circle. Then let your questions grow like vines. Don’t worry about making it look pretty. Just get the system out of your head and onto a page. You’ll be surprised what you see.
And if you need inspiration, the TestInsane repo is still a goldmine.
Thanks, Mike, for handing me that marker. And thanks to every messy tester who ever dared to draw what didn’t fit neatly into a table.
We may be in the age of automation, but some of our best tools are still hand-drawn.
—
Beau Brown
Testing in the real world: messy, human, worth it.
Testing Tester 
Leave a comment