A score needs to explain itself
The Reasoning Partner breaks the Board structure score into five factors: connectedness, relation clarity, supporting links, review signals, and reasoning depth.
Use these factors to inspect the graph. Supporting links are not verified evidence, and marked conflicts or dependency loops are places to examine rather than automatic proof of incorrect reasoning.
Best next move turns diagnosis into action
After the score is calculated, Canvio looks at the weakest factor and suggests one practical next step. If ideas are floating alone, it recommends connecting them. If links are unlabeled, it recommends clarifying relation meaning. If claims lack support, it recommends adding evidence.
The goal is not to grade the person using the board. The goal is to make a messy canvas easier to improve without forcing people to guess what to do next.
AI reads the board graph
The reasoning tools now pass a compact board summary to the assistant: visible nodes, relation labels, relation types, and map pin context where it exists. That helps the AI respond to the work on the board instead of giving a generic answer.
Challenge mode can look for weak assumptions. Socratic mode can ask better follow-up questions. Summary and article actions can use the current structure as source material.
Relations stay part of the canvas language
The board still uses the same direct interactions: quick-drag ports, typed labels, relation styles, and semantic relation types. The reasoning partner does not replace the canvas. It reads what people already built and helps them make it clearer.
This is important for everyday use. Students, teachers, researchers, and developers can keep thinking spatially while still getting useful feedback from the system.
Why this matters
Canvio is not trying to be another blank drawing surface. The larger idea is a visual knowledge workspace where objects, places, relations, and AI support the same thinking process.
A transparent score and a concrete next move bring that vision closer: the board can now help people understand their own structure, not only store it.