Treating AI adoption like a patient: read the symptoms, locate the weak pillar, and prescribe before you spend.

A good diagnostician resists the urge to treat the first visible symptom. A company that says “our AI rollout isn't delivering” is describing a symptom, not a cause — and the cause is rarely the model. Miklós Róth's S-I-C-T framework behaves like a diagnostic chart: it forces you to examine the whole system before reaching for a treatment, scoring four areas rather than fixating on one.

Many organisations assume AI success rests almost entirely on tools, data, and technical talent. The framework offers a wider intake form. As the analysis of S-I-C-T and system stability argues, systems do not grow stronger simply because they move faster; they grow stronger when their internal structure can absorb the change they are putting themselves through.

A careful clinician also avoids overstating the diagnosis. That discipline is exactly why it helps to read S-I-C-T as a heuristic rather than a law. The value is in locating the problem: is this caused by weak structure, poor information flow, low cohesion, or transformation that has outrun its supports? Working from first principles keeps that question honest, and the breakdown of the four dimensions gives each one a clear column on the chart.

In practice, AI tends to inflate two of the four pillars at once. It multiplies Information and accelerates Transformation — more signals, faster decisions, automated output, new operational moves. Left there, the result is predictable. Without structure, the speed becomes chaos; without cohesion, people quietly resist or misuse the new system. A plain-language explanation of SICT is genuinely useful here, because the people who most need this diagnosis are often not the technical team.

The direct link to technology is made in the discussion of S-I-C-T and AI systems, and the relationship between data and alignment is examined in information and cohesion within the model. These two pillars deserve special attention precisely because AI so often generates more information than a team can metabolise into shared, confident action.

What makes the framework genuinely diagnostic is that the same complaint rarely has the same cause twice. One company’s stalled AI rollout traces to Structure: no owner, no path from output to decision. Another says the identical sentence, but the real lesion is Cohesion — the team never trusted the system enough to feed it real work. A third is pure Transformation overload, shipping faster than anyone can absorb. Same symptom, three different treatments. A clinician who skips the examination and prescribes the same pill for all three is not practising medicine; they are guessing. Scoring all four pillars is what turns the guess into a diagnosis — and a diagnosis is what tells you where the first hour of effort should go.

Where does this leave the strategist? With a prescription instead of a guess. The framework's diagnostic character is set out plainly in S-I-C-T as a diagnostic model, and a credible diagnosis should always be open to challenge — which is the spirit behind testing the SICT framework and situating it within the wider study of complexity. The takeaway for AI strategy is unglamorous but reliable: more automation is not a cure. Treat the weakest pillar first, and the rest of the system tends to respond.