Governing the reasoning-to-action chain
Whitepapers for teams evaluating how autonomous AI should cross real-world effect boundaries: governing the chain from reasoning to action, producing audit-ready evidence for consequential action, and proving viability with buyer-reproducible telemetry. Policy decisions, fail-closed enforcement, and receipts are the substrate beneath.
Whitepapers explain strategy and architecture. Standards define conformance requirements.
The governed execution thesis.
Buyer Review Brief
v1.0
2026-07-07
Web + gated PDF
Executive, Technical, Compliance
- — reasoning-to-action thesis
- — audit-ready evidence for consequential action
- — Three Proofs: authority, causation, viability
- — BYOAI and Full Keon under one governed boundary
- — receipts and fail-closed enforcement as substrate
Current reading room publications.
AI Effects Need Receipts
v1.0
2026-07-04
Executives, Architects, Security Leaders, AI Platform Teams
- — pre-execution authorization proof
- — why autonomous systems must prove authorization before execution
- — mechanical floor for governed AI effects
- — relationship to the flagship reasoning-to-action thesis
Cortex: Evidentiary Memory Architecture
v1.0
2026-07-04
Platform Architects, Security Engineers, AI Infrastructure and Data Teams
- — verifiable causal evidence instead of mutable narrative state
- — Cortex memory preservation model
- — trust anchored in receipts rather than stories
- — technical architecture for evidentiary memory
Governing AI Ingress and Egress
v1.0
2026-07-04
Security Architects, Governance Owners
- — policy-bound inspection for agentic browsing and tool intake
- — outbound AI action controls before consequence
- — inspection requirements for untrusted content paths
- — governance obligations around ingress and egress
MCP Is Not Enough Without Policy-Bound Execution
v1.0
2026-07-04
Platform Engineers, Executives
- — why tool access alone is insufficient
- — identity, authorization, and receipts for MCP-connected tools
- — fail-closed governance at the execution boundary
- — production accountability for tool effects
Token Compression Without Losing Evidence
v1.0
2026-07-04
AI Platform Architects, Governance Owners
- — context cost reduction without destroying provenance
- — receipts, causal lineage, and auditability under compression
- — evidence-preserving compression pattern
- — offline verification implications for compressed context
Thoughts Are Free, Effects Are Governed
v1.0
2026-07-04
Architects, Risk Owners
- — governance model for multi-agent cognition
- — separation between free thought and governed execution
- — how much autonomy organizations should allow
- — multi-agent boundaries that prevent lawless execution
From the Hive to the Ledger
v1.0
2026-07-04
Platform Architects, AI Infrastructure Teams, Governance Owners
- — biological case for governed memory
- — how evolution's oldest collectives arrived at receipts and provenance
- — governed forgetting as a memory control surface
- — what hive memory patterns imply for enterprise AI memory
Choose the entry point that matches your review motion.
Start with the surface that matches your review role, then move into the governed execution thesis when you need the full argument.
For leaders evaluating operational AI liability, governance posture, and buyer risk.
For platform teams integrating governed execution with agents, MCP tools, policy gates, and receipts.
For teams mapping evidence trails, conformance statements, and audit-ready artifacts.
For teams understanding the category shift from advisory AI to operational AI.
Recommended reading path.
Use the sequence below if you are evaluating category fit, enforcement posture, and proof surfaces for the first time.
Whitepapers explain the argument. Standards evaluate the claim.
Whitepapers are explanatory. Standards are normative.
CAES and CPP define conformance language, requirements, and testable criteria. Use whitepapers to understand the argument. Use standards to evaluate claims.
Upcoming and planned papers.
These titles mark the research track without implying publication status beyond the label shown.
Evidence Packs and Portable AI Accountability
CAES Conformance and Effect Boundaries
OpenClaw and the Wild West of Exposed Agents
Need the enterprise version?
For evaluation teams, Keon can provide deployment architecture, evidence-pack examples, and governed execution walkthroughs scoped to your use case.