The tool is based on Quintessa's TESLA decision support software, which implements Evidence Support Logic (ESL), and an associated decision tree. Risks are classified into four categories: Capacity; Injectivity; Containment; and Monitoring. Each of these categories corresponds to a branch in the overall decision tree. The figure shows part of the containment branch. The top (root) node of this branch corresponds to the hypothesis that CO₂ containment can be demonstrated. This hypothesis is supported by other hypotheses at the next level of the tree. Where appropriate these hypotheses are then in turn supported by other hypotheses and so on, until a level of the tree is reached at which the hypotheses can be answered directly by evaluating site information. Evidence for and against each if the lowest level hypotheses, or 'leaf hypotheses' is judged independently and propagated through the tree by a numerical algorithm that combines evidence values for leaf hypotheses with the same parent hypothesis using user-defined 'sufficiencies' (effectively weights). Users of the tree may only input evidence values and may not change the tree's structure or the 'sufficiencies'. Consequently, the tool ensures that conversations among different project teams cover the same issues within each of these categories, to a similar level of detail. Users may embed supporting text and documentation within a tree, thereby producing an audit trail.