Joan — the Evidence Trail
A vertical iris trail runs through the picture with four repeated product needs: confidence scoring (it grades how sure it is), evidence grading (every claim carries its proof), identity matching (the same person recognized across systems), and owned memory (what it remembers stays yours, under your rules). On the left, five real systems — SignalClaw, Fabriq Platform, the predecessor Joan, Joan4U / Meeting Sidekick, and WIP — drawn as solid ink cards on cream paper, each connect to the trail. At the bottom the trail points into a dashed ochre box: Joan, shared context, memory, evidence, and review being assembled.
JOAN · THE BUILD PATTERN
The evidence trail
The same four product needs kept showing up across years of work — confidence, evidence, identity, and owned memory.
REPEATED ACROSS PRIOR BUILDS
SignalClaw
governed memory · confidence kernel
Fabriq Platform
identity survivorship across tenants
Joan — the predecessor
governed approvals · runtime truth
Joan4U · Meeting Sidekick
captured work → durable owned memory
WIP
connectors · carrying work back in
Confidence scoring
It grades how sure it is — about every single thing it keeps.
Evidence grading
Every claim carries its proof — confirmed, inferred, or unknown.
Identity matching
The same person, recognized across every system they touch.
Owned memory
What it remembers stays yours — held under your rules.
Joan — shared context, evidence, and review
the repeated pattern, assembled once, under one roof — being assembled now
Four needs kept coming back — now they point into one product layer the customer owns.
Joan
The evidence trail. Confidence scoring, evidence grading, identity matching, and owned memory kept reappearing across five separately built systems. Joan is where those lessons get assembled into shared context, evidence, and review — drawn dashed because that assembly is in progress.