How an engagement works

People come to us with a stalled system and a question. The system might be a clinical trial that isn't reading the way it should, a research program that has lost its direction, an organisation whose change initiative has gone quiet, or an AI tool whose effects on the surrounding cycle nobody has measured yet. The question is usually some version of what is going on, and what should we do?

An engagement with the practice is how we answer that together. It takes a shape we have arrived at slowly, through enough projects across enough different domains that the steps now feel reliable. Four moves, looped, never quite in clean sequence. We share them here so you know what to expect.

The wrong question, even asked beautifully, returns a wrong answer. So most of what we do at this stage is slow the question down. We ask what would change if we found the answer. We ask who is asking, and why now. We ask what would have to be true for the question itself to be the wrong one. Then together we rewrite it.

The asker is part of the question. A clinician asking why a trial failed is asking from a different position than the company that ran it, and a different position again than the patient who was in it. The same words point at different questions depending on where the asker stands and what they can act on. We try to make that position explicit at the start, because the answer that is useful from one position is often noise from another.

Every question lives in a system. Before we can localise where the system fails, we have to see how it works when it works.

The map we use is described on the dissipative systems page. It is general: the same map applies to a cell and to an organisation, because both are systems that consume energy, do work, and dissipate entropy in the same shape. When a new question arrives, we locate the substrate-specific version of it inside that general map. The map carries most of the structural work, which means we can spend our shared time on what is specific to your system.

With a map in hand, we can ask where the cycle stalls.

The same surface symptom often comes from failures at different positions. In cancer, a tumour that does not respond to anti-PD-1 can be a system stalled at the build phase, at the shipping phase, or at the encounter itself. The three patients look similar in the clinic and need different treatments. Treating them as the same problem is a common reason interventions fail. The same logic applies in other substrates: an organisation that has stopped renewing, a change program that has frozen, a trial whose biomarker isn't reading.

This step uses whatever measurement matches the resolution the question requires. Single-cell RNA when it is available. Bulk RNA when it is not. Cell-fraction proxies when the budget is smaller. Structured conversations and historical patterns when the system is an organisation. The measurement follows the question, not the other way around.

A localised failure narrows the field of useful interventions. Given where the failure sits, we ask which functions need to operate, and which interventions act at that position.

Two things shape the answer beyond the failure itself. The first is fit: an MCL-1 inhibitor and a CD40 agonist are both interventions in cancer biology, but they act at different positions and are interchangeable only when the failure sits where they both serve. The second is the asker's position, again. An intervention has to be one you can actually run, with the access, time, and authority you actually have. Anything else is a prescription that won't be filled.

Often the intervention already exists and is being applied to the wrong cases. Often no new drug or new program is needed at all. Sometimes the intervention is a measurement, a sequencing change, a conversation, a re-segmentation of who the program is for. We try to recommend the smallest move that lands.

An engagement can be a one-off structural read on a question you have, or a longer relationship where we sit alongside a research program, a trial team, or a change initiative as it unfolds. The practice is small. We take on a handful of engagements at a time so each one gets real attention.

Most engagements start with a conversation, no commitment, where we look at what you have and decide together whether the framework can help. If it can't, we will say so. That conversation is free.

The four moves loop. After an intervention has been tried, we measure again. Sometimes the failure has moved. Sometimes the system has resolved. Sometimes the question itself was wrong, and we go back to the beginning. We work on cancer, on populations under selection, on cycles, on organisations. The substrate changes. The procedure has stayed.