System 1 and System 2.
Most decisions are fast, automatic, and pattern-matched — not deliberate.
Kahneman's framing: System 1 is fast, intuitive, emotional. System 2 is slow, effortful, logical. We use System 2 far less than we think; most of life runs on autopilot.
That's why design beats argument. Change the default, the friction, or the framing, and you've changed the decision — without anyone having to think harder.
The applied version.
The two-system model is a simplification — Kahneman himself called it a "fiction useful as scaffolding" — but it works because it forces designers to stop assuming their users will think hard about anything. Almost no one thinks hard about almost anything, almost ever. Treat deliberation as the rare exception and design beats argument every time.
System 1 runs on four shortcuts: availability(what comes to mind easily must be true), representativeness(if it looks like the prototype, it is the prototype),anchoring (the first number sticks), andaffect (how I feel about it becomes what I think about it). These aren't bugs — they're the operating system. Your job as a designer is to set the defaults that System 1 will accept without waking System 2.
The applied insight: there are only three things a designer can actually change about a decision. (1) The default — what happens if no one decides. (2) The friction — how much effort each option requires. (3) The frame— the reference point against which the option is judged. Every behavior-change campaign that worked at scale moved one of those three. Every campaign that "raised awareness" without changing any of them did not.
This is also the most powerful lens on AI. Every UI choice in every LLM product is a default × friction × frame decision being made for billions of System 1 minds at once. The people who understand this will design the next decade. The people who don't will be designed by it.
Countries with opt-out organ donation have 80–90% participation. Opt-in countries: 10–20%. Same people, same values — different default. System 1 just ticks the box that's already ticked.
You'll ship one default change, one friction change, and one frame change inside your real work — and measure the lift in week 8.
- →The DFF Designer toolkit (Notion + Figma templates you keep).
- →Cognitive bias atlas with 88 biases mapped to the three levers.
- →Case work from healthcare, fintech, and HR tech engagements.
Which is the most reliable way to change behavior at scale?
Pick a behavior you want more of (or less of). Design a default change — not an argument — that would shift it.
Hint: Move the option, pre-check the box, change the friction, flip opt-in to opt-out.