Turn expert knowledge into structures that learners can genuinely understand — through conversation, not compilation. Claude serves as your Informed Assistant, probing for the derivations, analogies, and connections your knowledge contains.
You explain your knowledge. Claude probes for structure — asking how ideas derive from each other, whether derivations work in reverse, what models would demonstrate the connections, and where structural analogies exist across your domain.
As the conversation proceeds, topics and relationships are extracted and presented for your review. Confirmed elements flow into an interactive entailment graph that grows with each session — showing where the connections are and where gaps remain.
The Workbench continuously checks whether your domain is genuinely learnable — whether every topic can be reached from every other through derivation paths. Structural gaps are identified and the IA steers the next conversation toward filling them.
The Workbench implements Gordon Pask's Source/IA dialogue methodology. A subject matter expert (the Source) explains their knowledge; an Informed Assistant (the IA) applies five rules that transform a linear exposition into a cyclic, richly connected entailment mesh — the kind of structure that makes knowledge genuinely learnable rather than merely browsable.
The result is not a concept map (what goes with what) but an entailment mesh (how to get from any idea to any other idea, in both directions, through derivation and explanation).
The domains you build here become the terrain through which future learners navigate — supported by a conversational AI partner that adapts to how each learner thinks, assesses understanding through genuine dialogue, and creates conditions for learners to discover connections that neither they nor the expert anticipated.