What 49 Pragma conversations taught us about how leaders define their real problem

Pragma went live last week. Forty-nine conversations and 190 messages later, the data is already surfacing patterns that any change practitioner would recognise, but rarely sees quantified.

The first thing people present is almost never the real problem.

In roughly half of all conversations, the topic shifted between the opening message and the final exchange. Someone arrived asking about communication strategy and ended up working through whether they had genuinely committed to the change themselves. Someone framed it as a resourcing problem and discovered, three exchanges in, that it was a sequencing problem.

This mirrors what happens in consulting engagements. The real question tends to emerge in the third or fourth meeting, not the first. Pragma just compressed that into four messages.

Generic questions produce zero deep engagement. Literally zero.

Openers that named a specific person or situation reached deep engagement at 58%. Generic how-to questions reached it at 0%.

The practical implication extends well beyond an AI tool: if you cannot name a person in your change problem, you are probably not ready to work on it yet. A generic question is a search query. A specific situation is the beginning of a real conversation.

What this means for how Pragma works

These patterns are shaping how we develop Pragma as a change advisory tool. The conversation design is built to move past the presenting question, not to answer it faster, but to help leaders find the question underneath it.

If you are leading a complex change right now, or sitting with a problem you have not quite been able to name, try Pragma. Bring a real situation, not a hypothetical. The data says that is where the value starts.

Mark's full analysis, including the bimodal trust pattern and why he is considering removing the template prompts, is on markwinter.com.au.