Answers are grounded in this paper and cite their sources.
Enterprise adoption of agentic systems is often framed as a problem of agent reasoning, tool use, retrieval, or memory. This paper investigates an additional architectural problem: preparing heterogeneous enterprise knowledge into task-appropriate representations that agents can interpret, reuse, and govern. We propose Cognitive Preparation as an explicit architectural lifecycle between authoritative enterprise knowledge and agent execution, with a Cognitive Preparation Compiler (CPC) coordinating representation decisions, compilation, validation, provenance binding, temporal constraints, and publication through a reusable Cognitive Substrate. The contribution is positioned at the architectural and methodological level; no novelty is claimed for ontologies, symbolic predicates, memory mechanisms, provenance, temporal models, or knowledge compilation individually. Recent 2026 preprints on precomputed agent-memory views and enterprise context engineering further narrow the novelty boundary and are treated as close, non-peer-reviewed prior art. We report a staged prototype evaluation implemented in Ver.Test9 on a Ver.Test8 runtime foundation. Deterministic tests examine selected lifecycle controls, while exploratory LLM-assisted evaluations (T17–T22) examine controlled reuse of synthetic operational knowledge through deterministic retrieval. The results provide implementation-level feasibility evidence and preliminary observations of scoped procedural knowledge reuse, including provider- and operation-specific applicability, temporal validity, and version precedence. They do not establish general performance superiority, validated adaptive representation selection, transfer between genuinely independent agents, autonomous learning, production readiness, or economic break-even. The five research questions are retained, with preliminary support strongest for controlled knowledge reuse; preparation effectiveness and evolution receive partial evidence, while adaptive representation selection and reuse economics remain open.
No takes yet. Share an insight, caveat, or question.
Edison Sira Carreño (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: