PayGraph

The model

From social connections to financial ones

The Social Graph connected people, friends, and communities so software could understand who knows whom. The PayGraph does the same for money: accounts, merchants, devices, household, work, rewards, and goals—centered on identity, not on a single institution's ledger.

Social Graph

Who you know

Profiles, friendships, follows, and shared content—relationships that make social software useful.

PayGraph

How your financial life connects

Payments, accounts, commerce, permissions, and intent—relationships that make financial software useful without raw data dumps.

What it is

  • A living map of permissioned financial and commerce connections around a person.
  • Facts and inferences kept distinct—with confidence, policy, and a path to correct them.
  • Intelligence that sits above banks and networks without replacing their systems of record.

What it is not

  • Not a replacement for a bank ledger or payment network.
  • Not a pile of every raw financial record in one place.
  • Not surveillance, a universal consumer score, or selling data.

Fundamental objects

The vocabulary of the graph

IdentityVerified person or authorized organization
NodeEntity or concept—account, merchant, device, goal
EdgeTyped relationship—owns, funds, trusts, authorized by
EventSomething that occurred—auth, purchase, consent grant
SignalObserved context—device reputation, velocity, location
ClaimGoverned statement usable by another service
PolicyPermissions and restrictions by purpose and scope
ScoreBounded assessment with lineage and confidence
InsightHigher-order interpretation—not a raw fact
ActionApproved output—approve, step up, recommend, explain

Knowledge classes

Facts and inferences stay distinct

A critical design rule: the PayGraph must never present an inference as a fact. Every derived attribute retains provenance, confidence, freshness, policy, and a path to correction.

ClassExample
DeclaredUser states a goal to save for a home
VerifiedInstitution verifies account ownership
ObservedA payment authorization occurs
DerivedDining spend rises on business-travel weeks
InferredUser may be planning an international trip
AggregatedAnonymous cohort shows a seasonal pattern

Stack

From connected systems to experiences

Connected systems

Banks, networks, processors, wallets, merchants, loyalty, identity providers

Event ingestion

Auth, schema validation, source attribution, deduplication, replay protection

Tokenization & normalization

Sensitive IDs replaced, merchants standardized, lineage preserved

Identity & consent

Resolution, roles, purpose, revocation, policy versioning

Graph & feature stores

Time-aware nodes, edges, projections, embeddings, snapshots

Intelligence services

Scoring, anomaly, recommendation, eligibility, explanation, simulation

Developer & agent APIs

Claims, GraphQL/domain APIs, SDKs, sandbox, webhooks, metering

Experiences

mCards, Neobankify, partner apps, portals, support tools