Thirty years of financial semantics.
PAIX — the Protocol for Asset Information eXchange — a purpose-built, human-readable semantic model for institutional finance, extensible by your analysts and by AI.
PAIX, our personalizable ontological framework for investors, was born when Lloyd Oestreicher realized that organizing data like shipping containers offers huge functional advantage.
A model of the business, not another schema.
PAIX is a semantic model architecture for institutional finance — an ontology-backed application data layer designed to be customized by users. It directs the management of data by describing the objects a firm deals in (positions, instruments, clients, mandates), their attributes, and the relationships between them, so meaning travels with every record. We created the category and have refined PAIX for three decades.
A source of harmony, not a golden source.
It is battle-tested — proven over decades on trillions of dollars in assets and well over $100T in transactions, plus all related reference and other data. PAIX harmonizes many sources of truth rather than forcing incompatible apps, processes, and business lines into one awkward framework: broader, deeper, and more reliable than anything built from scratch.
What sets PAIX apart.
Harmonize the information, systems, and people across your organization on a semantic architecture that molds to you — and generate more revenue at lower capex and unit cost, tomorrow, not next year.
Extended in place, not rebuilt.
Extend the model to your own objects and definitions in place — individually, transparently, and fast — by an analyst or with AI assistance. More rapidly and easily customizable than any generic ontology tool.
Decades, trillions, no breaches.
Thirty years in production across trillions in assets and $100T+ in transactions, reconciled to 99.9997% accuracy — broader, deeper, and more reliable than a build-your-own model, with an unbroken security record.
Reasoning, already attached.
ALPHA reasoning is already wired in — Northfield risk and optimization, Market76 personalization and client profiling — with no translation layer between meaning and math. And because the model extends and ingests new sources automatically, any other reasoning you want to attach comes in the same way.
Six products.
One record.
Each does one job to a standard that stands on its own — and shares a single semantic model, so they interoperate out of the box.