The 11 transversal AI capabilities, arranged under the four canonical sub-domains: knowledge quality (gap detection, grounding, temporal validity), symbolic inference (cascade, OWL materialization), neural extraction (entity recognition, sovereign transcription) and assistance (context, explanations, local inference, schema-constrained LLM). Four shipped, four partial, three gap. Symbolic always wins; neural is subordinate. Each card carries its real status.
Each column is a canonical sub-domain. Click a capability to open it in the referential.
Evidence-driven intelligence runs symbolic-first: the cascade, OWL materialization, gap detection and schema-constrained extraction are shipped and always authoritative. Grounding, explanations, context and entity recognition are partial; temporal validity, sovereign transcription and local inference are the honest gaps. Statuses derived from the referential (capabilities.html cap-NS-1..11, 2026-07-09).