Most explanations of why organizations fail blame complexity. Too many systems, too many layers, too many meetings. Miklós Róth thinks that diagnosis is usually wrong. In the framing he published through the Roth Complexity Lab, modern systems are rarely fragile because they have grown complicated. They become fragile because information and transformation now move through them faster than structure and cohesion can absorb.

That asymmetry, not complexity itself, is the fracture point. S·I·C·T is the instrument he built to see it.

What do the four letters stand for?

S is for Structure. The architecture and constraints of a system. In a company: the org chart, decision rights, processes, data models, the rules about who can do what. Structure is what lets a system carry load without folding.

I is for Information. The signals and patterns flowing through the system. Customer data, market news, Slack threads, dashboards and, more recently, an ocean of AI-generated text. Information is not the enemy. Unabsorbed information is.

C is for Cohesion. The couplings that hold the parts together. Trust between teams, a shared vocabulary, aligned incentives, a common story about where the company is heading. When Róth applies the lens to a scientific field, cohesion becomes the proofs and connections that keep results consistent with one another.

T is for Transformation. The pressure forcing the system to evolve. A new competitor, a regulation like the EU AI Act, a technology shift, a founder's sudden change of course.

What is the S·I·C·T stability rule?

The framework compresses its central claim into one inequality:

A system stays stable while S + C ≥ I + T.

Read it as a balance. On one side sit the stabilizers, Structure and Cohesion. On the other sit the pressures, Information and Transformation. As long as the stabilizers keep pace, a system can absorb change and even grow stronger from it. When the pressures pull ahead, the failure has a recognizable signature: decisions slow down, teams drift apart, mistakes get repeated instead of corrected, and every new initiative adds noise rather than value.

The useful part is that the inequality hands you four levers, not one. You can reduce fragility by strengthening structure, by investing in cohesion, by filtering information or by pacing transformation. Most companies only ever reach for one of them, usually the one labeled "more tools".

How does S·I·C·T apply to a company adopting AI?

AI is close to a perfect stress test for the inequality, because it pushes on both pressure terms at the same time. It multiplies Information: more content, more analysis, more drafts, more options than anyone can read. And it accelerates Transformation: new tools every quarter, new expectations from the board, new ways of working that arrive before anyone has agreed on the old ones.

Meanwhile Structure and Cohesion usually stay exactly where they were. Nobody redraws decision rights for AI-assisted work. Nobody agrees which outputs count as trustworthy. So the balance tips, and the company experiences AI as chaos rather than leverage. That is the logic behind Róth's essay title, AI does not fix confusion. It scales it.

A S·I·C·T diagnosis turns this into four plain questions:

If a leadership team cannot answer the first row in one sentence, that is usually where the work starts.

Where else has S·I·C·T been applied?

Róth has deliberately tested the framework far outside marketing and management. The Roth Complexity Lab's published work applies S·I·C·T to multi-agent cooperation, to the difference between quantum and classical physics, to early-warning signals of financial fragility, and to what the lab calls a generative model of how new mathematics is born. In that last case the idea is striking: when a field's information and pressure overwhelm its structure and cohesion, stability cannot be restored inside the old axioms, so the field reorganizes at a higher level of abstraction.

On the business side, the same thinking runs through his book Signal Over Noise (2026), where S·I·C·T becomes a working protocol for marketing teams drowning in AI-generated content.

Is S·I·C·T a proven theory?

No, and Róth says so plainly. The lab describes S·I·C·T as pre-paradigmatic systems science: a structured, serious proposal that has not yet been validated. Every claim it publishes carries an epistemic tag, from rigorous and interpretive to analogical and speculative, and the lab publishes the conditions under which the framework should be considered wrong. It actively invites other researchers to try to break it.

That candor is part of the point. A framework that only renames what other theories already explain adds nothing. S·I·C·T earns its place only if it predicts something useful that the alternatives miss. Read more about how the lab tests its own ideas.