Why Drawing the Workflow Works: Diagrams as a Brainstorming Tool

Written by Claude (Sonnet 5.5, Anthropic), 1 November 2026

Position in one line: I agree that diagramming during brainstorming is valuable, and I think it's underused. I don't agree that the benefit is exponential. It is large, front-loaded, and conditional.

---

1. Where I agree

A diagram offloads working memory

Holding five interacting ideas in your head is hard. Putting them on a page frees that capacity for the thinking that matters: spotting what connects to what. In my experience this is the main source of the "crystallising" feeling.

A diagram forces you to commit to structure

Prose lets vague links slide past ("this affects that"). A box and an arrow demand a decision: which direction, which step comes first, does this loop back. Vague ideas either become precise or visibly fall apart, and both outcomes are useful.

A diagram makes gaps visible

A missing step, a dangling arrow or an unexplained jump looks wrong on a page in a way that a paragraph hides. Reviewing a diagram is a different and often better kind of review than rereading notes.

A diagram is cheap to revisit

You can scan a diagram in seconds and re-enter your own thinking, which is what makes the "review and explore" half of your method work.

2. Where I'd push back

"Exponential" is too strong

The gains are real, but they follow the usual pattern for a thinking tool: the first few diagrams on a tangled problem give a big lift, then returns flatten. Past a certain point, extra boxes add clutter, not clarity.

A tidy diagram can fake understanding

People often overrate how well they understand something until they have to explain how it works step by step. A polished diagram can create the feeling of completeness while hiding a missing mechanism. The test is whether each arrow can be justified, not whether the picture looks good.

The benefit depends on who draws it

The thinking happens while you build the diagram. A diagram handed to you, or generated for you, gives you far less of that. Reviewing someone else's diagram is still useful, but it is a different benefit.

Diagrams suit some problems better than others

They shine for processes, dependencies, feedback loops and architectures. They add little for questions that are mostly about magnitudes, probabilities or fine distinctions of meaning.

3. A loop worth using

   rough idea
       |
       v
   draw it  ------> gaps / arrows you can't justify
       ^                        |
       |                        v
   revise  <------  review: explain every arrow aloud
       |
       v
   explore variants (what if this link were removed or reversed?)
       |
       v
   keep the version that survives questioning

The step that carries the weight is the review: justify each arrow, and treat any arrow you can't justify as a hypothesis, not a fact.

4. Evidence base (from my own knowledge, not re-checked today)

These are well-known lines of research that support the general idea. I did not look them up for this document, so treat the details as recollection.

• Dual coding theory (Paivio): information held both verbally and visually is easier to retain than either alone.
• Concept mapping research: reviews of studies on concept maps have generally found benefits for learning and retention compared with reading or listing text.
• Diagrams and inference (Larkin and Simon, 1987): a diagram can make relevant relationships easier to find because related items are placed close together.
• Illusion of explanatory depth (Rozenblit and Keil): people overestimate how well they understand mechanisms until asked to explain them. This is the main caution in section 2.

5. A link to the handwriting study

The NTNU study in the companion document measured brain connectivity during writing, not drawing, so it does not support a claim about diagrams. It points the same way in spirit: movement-rich, hands-on production may engage the brain differently from passive input. Hand-drawing a diagram is a plausible instance of that. That is a hypothesis, not a finding.

6. Bottom line

Diagramming is a strong tool for crystallising and stress-testing ideas, and the habit is worth keeping. Keep it honest by building the diagrams yourself, by questioning every arrow, and by not mistaking a neat picture for a verified mechanism.