Research & examples
This library contains active studies, working papers, technical notes, and synthetic examples. Unfinished work is labelled unfinished. Synthetic work is labelled synthetic. Public recommendations are not presented as client outcomes.
Current research
The priority is original research, controlled methods, worked examples, and reusable frameworks rather than a high-volume stream of generic AI commentary.
A controlled study testing whether provenance, uncertainty, disconfirmation, and final claim-audit controls reduce unsupported claims in AI-assisted research without winning by saying less.
Open the study recordA structured explanation of intent graphs across semantic routing, task-dependency planning, skill composition, and commercial search journeys, with a MainFrame implementation lens.
Read the technical noteA research program mapping existing work on false-completion detection, appropriate abstention when evidence is absent, and evaluator self-calibration against blind human gold.
Method walkthroughs
These pages show method and deliverable shape without borrowing client credibility or publishing peer materials.
Design-sprint method on a synthetic OOS investigation: reconstruction test, record chain, R1-R8, gaps, and design-only plan.
Open walkthroughChange path from published amendment to cutover: authoritative records, failure signals, owners, evidence.
Open walkthroughHow claims are separated from evidence, with a full synthetic buyer-readiness structure.
Open walkthroughPublication controls
Research and examples identify the author, status, version, evidence or data basis, method, intended use, limitations, and availability of code or supporting material where applicable. Material changes should produce a new version rather than a silent rewrite.
The Evidence Intelligence Research Sprint converts a consequential question into a controlled source register, analysis, decision memo, appendix, and briefing.