Before I send an analysis memo to counsel or publish it here, I run through the same checklist. It keeps me honest and prevents surprises during interviews.

1. Scope lock. Confirm the question I’m answering (invalidity search, competitive landscaping, whitespace study) and restate it at the top of the document.
2. Dataset hygiene. Snapshot the data sources, API versions, and filters used to build the corpus. If I relied on AI summaries, I paste the prompt + parameters so the work can be reproduced.
3. Claim mapping. Highlight which claim elements each reference covers. I often include a table so legal teams can port it into their own templates.
4. Risk notes. Flag assumptions, chemistry-specific caveats, and manufacturing realities I learned from industry.
5. Automation hooks. Whenever a manual step takes longer than ~10 minutes, I log it as a scripting opportunity. Those notes feed future posts.

The list evolves, but the principle stands: every deliverable should be verifiable, auditable, and useful to both scientists and attorneys.