AI second brain as a personal assistant

Challenge

Content lived in different files - locally, on the phone, or in the cloud. Re-finding decisions and source material cost time every week, and no single place held the full picture.

Intervention

Built a personal second brain as an LLM markdown wiki - immutable raw sources, model-owned concept pages, and a schema so the assistant maintains links and an index instead of dumping answers into chat.

Measurement

Compared time-to-answer for recurring questions (prior decisions, entity facts, linked sources) before and after the wiki was in daily use.

Result

Answers came from linked wiki pages with citations back to sources, not from hunting across devices and folders. Time-to-answer and knowledge reuse improved versus the before window as the wiki compounded.

Next step

Documented the pattern so similar second-brain setups can be reused for delivery leads and small specialist teams.

Wiki

Knowledge shape

Before / after

Comparison

Faster reuse

Time-to-answer

Results depend on workflow, role, baseline and measurement method.

Discuss an engagement