Second Memory

Second Memory is a guided Personal AI Context Audit for people who use AI regularly.

Over time, an AI can build a working portrait of your projects, priorities, preferences, decisions, and ways of working. Much of that context may be accurate while still being outdated, incomplete, overweighted, or arranged around the wrong priorities.

Second Memory helps you inspect that portrait.

Using either a quick Express Audit or a more rigorous Deep Audit, you work with your own AI to look for five common context problems:

Wrong Geometry — correct facts, wrong hierarchy
Uneven Forgetting — important context receives too little weight
Recursive Influence — repeated descriptions gain too much authority
Commitment Inflation — discussion is mistaken for commitment
Outdated Context — old information still shapes current advice

You then decide what your AI should preserve, qualify, reduce, retire, ask about, keep private, or leave unresolved.

By the end of the workbook, you leave with three practical outputs:

Verified Current Portrait
A concise description of the context your AI should currently use when working with you.

Second Memory Findings
A record of any material problems discovered during the audit.

Second Memory Charter
Your rules for how approved context should be used going forward.

The workbook also includes an optional Current Context File, a Live Test to see whether the audit actually improved your AI’s response, and a Five-Minute Second Memory Check for future maintenance.

You do not need technical knowledge. You do not need to expose your whole life. And the goal is not maximum memory.

The goal is useful, current, and human-authorised context.

Format: 33-page digital workbook
Includes: Express Audit, four-module Deep Audit, governance and synthesis, Live Test, and maintenance check