Revelo

The ANU School of Cybernetics needed an intuitive way to teach systems mapping within time-constrained workshops. Traditional methods are often too complex and time-consuming for students to grasp quickly, with mapping sessions spread over hours to days. We created Revelo to streamline this process so a group could collaboratively build a systems map in under an hour. Additionally, understanding ongoing dynamics from a systems map is hard for newcomers, so we strived to make the resulting maps more understandable.

To create maps collaboratively, we use participants' mobile devices to collect their understanding of the system. Participants contribute and vote on the importance of different variables and how those variables relate to each other. We then collate these votes into the systems map, identifying areas of conflicting understanding for discussion in the workshop. We produce a map that models the system dynamics from these variables and relationships. This process, typically done over hours on pen and paper and often dominated by particular viewpoints, takes less than 20 minutes while giving every participant representation.

A key design challenge that emerged during early testing was how to rapidly find relationships without the frustration of evaluating every possible combination separately. I designed an optimised voting interface that significantly accelerated the process by highlighting a central variable at the top while allowing users to label its relationships with all other variables on a single screen. This solution maintained the flexibility of traditional pen-and-paper approaches while dramatically improving efficiency and engagement.

Once users create a system map, it is only valuable if they can find meaningful insights and understand the implications of their modelling. Revelo simplifies this process by automatically identifying all feedback loops that define the system's dynamics and describing those in natural language.

By analysing how variables interconnect, we mathematically determine whether these loops are stable, reinforcing, or likely to collapse. This data provides a factual foundation for our LLM implementation, minimising AI hallucinations while giving workshop participants clear insights into emergent behaviours. By automating both the identification and explanation of feedback loops, Revelo makes complex systems thinking accessible in ways traditional mapping methods cannot.