The Transparent AI in Education framework (Appendix B of The Practitioner) asks students and practitioners to maintain a lineage document — a complete record of how AI-assisted work was produced. The first time through, this feels like significant overhead. That is intentional: doing it manually builds the habit of thinking that makes the practice valuable.
Once the habit is established, the manual process should be replaced with tooling. This skill is about that transition: how to go from deliberate manual logging to an automated workflow that captures your lineage as a byproduct of how you work.
Workflow
Three Levels of Automation
The transition from manual to automatic happens in stages. Start at Level 1 and move up as the habit solidifies.
The simplest upgrade from fully manual. At the end of each work session, submit a single prompt to your AI:
"Summarise this session as a lineage entry. Include: the original intent, the prompts I used (summarised), the AI contributions, the points where I overrode or redirected the output, and the final decision made. Format as a Markdown section I can append to my project log."
This requires the AI to have access to the session context — either because you are in the same conversation, or because you paste in the relevant exchanges. It adds two to three minutes to the end of a session and produces a structured log entry without manual construction.
Best for: People new to lineage logging who want to reduce friction without changing their workflow significantly.
The Log Structure
Regardless of level, the output should match the lineage format from Appendix B. A lean version:
## Session Log — [Date] — [Project]
### Intent
[What this session was trying to accomplish]
### AI Contributions
[Summary of what the AI produced or suggested — verbatim for key prompts,
summarised for routine exchanges]
### Key Decisions
[Where human judgement shaped the output — what was accepted and why]
### Overrides
[Where AI output was rejected or significantly changed, with brief rationale]
### Output
[What was produced and its status: draft / reviewed / committed]
Connecting to Appendix B
The Appendix B framework distinguishes Phase 1 (research and refinement) from Phase 2 (the build). Automated lineage logging applies to both — the structure is the same, the session types differ.
For educational contexts, the automated log is the evidence of genuine intellectual engagement. A log generated by committing the session in real time, rather than reconstructed from memory after the fact, is more reliable and more honest than one written retrospectively.
For professional contexts, it is the audit trail. Employers and clients who need to understand how AI-assisted outputs were produced can read the log and see exactly what happened.
The goal in both contexts is the same: a practitioner who can show their working — automatically, without friction — is a practitioner who is trusted.
From Appendix B: Transparent AI Use in Education — The Practitioner (Phil Rust, 2026). Part of the Practitioner companion resources.