Pursula

Pursula builds AI decision engines: systems that turn real-world signal into structured data, record the decisions made in the moment, and use that record to model what should happen next.

Most technology digitises the task and loses the lesson. Data piles up on dashboards while the link between a decision, the conditions it was made under, and what actually happened quietly disappears — so the same cost of learning gets paid again every time the tools change. We call the alternative decision memory: keeping that link on purpose.

Structurally, that makes Pursula an engine company rather than a single application: one system for turning media, social, and behavioural signal into decisions, built to be pointed at whatever domain needs it next. We organise ourselves the same way — around outcomes, not job titles — and examine failure out loud instead of hiding it. A manual fix that keeps showing up in the field is a feature waiting to be built.

AI decision engines that turn real signal into real outcomes.

We believe systems should learn from people, not about them. A system that makes decisions has to be able to explain them, and the data, the history, and the ownership of the process stay with the person whose work it is.

Real problems. Measurable outcomes. Decision memory that persists.