artifact-pyramids

You can organize AI research outputs by mapping them into a three-layer artifact pyramid: L1 Summary, L2 Analysis, and L3 Dossiers. This structure reduces cognitive load by allowing different agents to consume only the depth of context they need.

Progressive disclosure structures AI outputs into hierarchical layers, enabling role-appropriate access. Product managers, analysts, and validators navigate from an L1 summary down to L3 dossiers via explicit SOURCES links, accessing detailed evidence only when required.

You route multi-agent pipeline outputs by assigning roles to specific pyramid layers. Product managers consume L1 summaries, analysts use L2 analysis, and validators audit L3 dossiers, ensuring precise agent routing and targeted consumption across the pipeline.

Yes, you can reuse and audit research dossiers by maintaining a flat L3 dossier library. The system supports an amending lifecycle that creates an amended root pyramid per epoch, allowing evidence reuse and auditing across multiple pipeline runs.

No specific pipeline frameworks are required. Artifact pyramids apply across various profiles and pipelines, enabling you to enforce structured SOURCES navigation and amending lifecycles without external framework dependencies.