The traditional customer journey is drawn as a funnel: awareness, consideration, conversion, retention. But real relationships are not linear. A payment failure, a support escalation, a four-year gap, and a sudden return all mean different things depending on the history that came before them.
A process-native view treats the relationship as a statechart: nested lifecycle states, parallel regions of interest and trust, history markers that restore context, and transitions guarded by events. The same click means something different depending on where the relationship already is.
Interactive statechart
Follow the synthetic customer journey below. Use the timeline and play controls to walk through episodes, then inspect the proposed statechart structure, the event log, and the runtime induction path.
From event logs to an executable runtime
The structure shown above is not just a visualization. It can be induced from real behavioral evidence and then compiled into a durable runtime:
- Observe CRM, billing, support, product, and message events.
- Discover repeated episodes, boundaries, parallel concerns, and outcomes.
- Propose LLM-assisted state names, hierarchy, transitions, guards, and actions.
- Approve + run the human-reviewed model as the relationship runtime.
This is the same compile-then-execute pattern behind RunTime: the model becomes the owned asset, not the frontier-model call.