The contact book: who's who, without the upkeep
Your conversations mention people: colleagues, clients, one-time callers. KISMO keeps the register for you: one card per person, fed note after note.
A book that fills itself
With every analyzed note, the people present or mentioned join your book: identity, role, details heard in the conversation. Recognition relies on strong signals first, phone number and email. Someone who is only mentioned, but carries useful information, gets a card too.
One card per person, one single truth
The card gathers what your notes know about someone: role, company, attributes collected across conversations, each tied to the notes where it appears. Fix a name once: it updates in every note, export and share. What you validate is never overwritten by AI, and what AI detected stays flagged as such. The name display format you choose applies everywhere.
Duplicates show up, you decide the merge
The same person keeps coming back under three spellings? KISMO groups the cards that probably describe one person and shows you the resulting card before merging. You arbitrate field by field, or dismiss the match in one click. Absorbed cards go to the trash, recoverable. Names distorted by transcription get caught up: the book learns from your corrections and matches spellings that sound alike, and never merges on sound alone.
The network draws itself
Who works with whom? The book's graph offers two readings: connections, when people appear in the same notes, and relationships, extracted from what was said (colleague of, reports to, manages). Path mode links two people and shows through which notes they touch.
Search the way you speak
"Who's the tax expert?" The book's search understands what the question means: it covers names, roles and card attributes. The book also answers in the global search, everywhere in your account.
The people in a note, under control
Every note lists its participants and the people mentioned. You confirm uncertain identifications, add someone from the book, replace one person with another or swap two mixed-up attributions; the wording adjusts on its own. In the editor, the @ mention picks straight from the book.
A history that loses nothing
Each card keeps its versions: original state, merges, corrections. Restoring a version erases nothing, the current state stays in the history.
The book feeds the rest of KISMO: the assistant knows who's who when it answers (AI assistant), and cards follow your notes in the workspace.