Learn it → Apply it → Save as Skill
A checklist that works beautifully on a known process can fail badly on a problem nobody has seen before, and a careful analysis can waste precious time in a crisis that needs action first. The Cynefin Framework sorts a problem into the kind of territory it actually sits in, so you reach for a method that fits instead of the one you already know.
The Dr. Wallace’s Favorite Mental Models series is designed to turn useful ways of thinking into reusable tools. Each model gives you a fast visual overview, deeper interactive learning, practical application exercises and an AI-ready Markdown skill you can save for future use.
Instead of merely reading about mental models, you can gradually build your own AI skills library: a collection of portable reasoning skills that can help ChatGPT, Claude, Grok and other AI assistants apply the mental models you value.
Some problems have a known fix. Some need expert analysis. Some only make sense after you've acted and watched what happened. And some need you to act immediately, with no time to analyze at all. Cynefin names these four territories so you can match your method to the one you're actually in.
Important: none of these problems are less real than the others. What changes is how visible the cause-and-effect pattern is before you act, and that changes which method actually has a chance of working.
The model is most useful at the start of a decision, project or crisis, before you've locked in an approach, when getting the category right determines whether your method will even work.
The Cynefin Framework doesn't tell you the answer to a problem. It tells you which family of methods actually has a chance of working on it.
Gather enough signal to judge whether cause and effect are obvious, analyzable, only visible in hindsight, or genuinely unclear right now.
Place the situation into one of four domains: Clear, Complicated, Complex or Chaotic, or flag it as Confused if you honestly can't tell yet.
Match your method to the domain: documented procedure in Clear, expert analysis in Complicated, safe-to-fail probes in Complex, decisive action first in Chaotic.
Re-sense regularly. Problems migrate between domains as conditions change, and yesterday's right method can quietly become today's wrong one.
Open a use case to see what the model changes.
Resist pulling out the standard troubleshooting checklist if the failure pattern doesn't match anything you've seen. Treat it as Complex: stabilize first, then run small, reversible diagnostic probes rather than guessing at a single root cause.
A market with no stable pattern is Complex, not Complicated. Run small, reversible experiments and watch what the market actually tells you, rather than commissioning one big analysis and committing to its conclusion.
Routine, well-understood steps belong in the Clear domain with a documented procedure. Steps where experienced people still disagree belong in Complicated, with room for judgment. Forcing both into one rigid script creates friction in the Clear cases and bad outcomes in the Complicated ones.
A task with an agreed procedure and a visible right answer is Clear or Complicated, and a reasonable automation target. A task where the "right" output is still debated among skilled people is Complex, and shipping it as if it were settled produces confidently wrong results at scale.
This usually starts Chaotic: stop the immediate damage with a fast, contained action, even an imperfect one, before you have time for careful message-testing. Move to Complex once the immediate danger is contained.
Early-stage, fast-moving technology usually sits in the Complex domain: write provisional rules, watch what actually happens, and revise, rather than writing one comprehensive policy as if the long-term effects were already knowable.
This is an illustrative example designed to teach the method.
Expand for where the term came from and how it has evolved since 1999.
Dave Snowden developed it in 1999 while working at IBM Global Services, refining it further as European director of IBM's Institute of Knowledge Management. He founded the IBM Cynefin Centre for Organizational Complexity in 2002, which became independent of IBM in 2004.
It is a Welsh word with no single English equivalent, often translated as "habitat" or "place of multiple belongings." Snowden chose it to suggest that people's understanding of a situation is shaped by many factors, cultural, historical and experiential, that they are only partly aware of.
Yes. Early formulations around 2003 used Known, Knowable, Complex, Chaotic and Disorder. From around 2014 onward, Snowden renamed them Clear, Complicated, Complex, Chaotic and Confusion, aiming for labels that were harder to misread as a ranking of difficulty rather than a description of how visible a pattern is.
It is rooted in complexity science rather than a single controlled study, and was popularized through the November 2007 Harvard Business Review article "A Leader's Framework for Decision Making" by Snowden and Mary Boone. It is widely used in consulting, government and technology organizations as a sense-making tool; critics note that some of its terminology is hard to apply consistently and that it has had limited formal empirical testing.
Assuming enough expert analysis will eventually reveal the one right answer, when the situation is actually emergent and only becomes clear through small experiments, not more analysis.
Applying the standard procedure confidently to a situation that doesn't actually match any known pattern can make things worse, not just fail to help.
Calling a meeting to assess options while a crisis is actively unfolding, when the moment calls for a fast, stabilizing action first and analysis second.
A process that was genuinely Clear can drift into Chaotic through neglect or changing conditions; the sort needs revisiting, not deciding once and filing away.
The Complex domain still has a method: small, safe-to-fail probes with real monitoring, not random action with no attempt to learn from it.
"We've dealt with something like this before" is a feeling, not a diagnosis. Check whether the current cause-and-effect pattern is actually visible before assuming the domain hasn't changed.
the decision is low-stakes and easily reversible. A full domain diagnosis costs time that a quick, correctable choice doesn't need.
a way to publicly blame someone for a "chaotic" situation. The framework is meant to choose a response method, not assign fault for conditions nobody fully controlled.
the domain is already decided. Cynefin picks the right family of approach; it doesn't replace the detailed expertise, analysis or experiment design you still need once you know which domain you're in.
This browser tool structures your thinking. It does not send your information anywhere or perform AI analysis.
Save as Skill converts the model into portable Markdown instructions that teach an AI when and how to apply the Cynefin Framework, what questions to ask, what evidence to seek, what mistakes to avoid and how to structure its analysis.
Save the skills you find useful and organize them into your own reusable library. The Markdown is intentionally model-neutral, so it can be used wherever an AI system accepts uploaded files, project knowledge, custom instructions or reusable text context. Platform capabilities differ, so this is a portable skill file rather than a claim of one-click installation.
A risk matrix ranks problems by likelihood and impact on fixed axes. Cynefin sorts problems by how visible their cause-and-effect pattern is, which changes the method you use to solve them, not just how much attention they deserve.
Yes, and this is common. A product launch might have a Clear logistics checklist, a Complicated technical build, and a genuinely Complex question of how customers will actually respond, all running at the same time.
In Complicated situations, enough expertise and analysis will reveal the right answer in advance, like diagnosing a car engine. In Complex situations, the right answer only becomes visible after you act and observe the result, like predicting how a market will respond to a new product.
It names the honest state of not yet knowing which domain you're in, often because different people are each assuming a different one. Naming it is itself useful: it prompts breaking the situation into smaller pieces that can be sorted individually, rather than guessing.