Open PortfolioOpen Portfolio.
BYOC / InspectableInspectable BYOC boundary — buyer keeps IdP and approved storage; inference runs over bounded aggregate context.

Edge Ingestion vs Client-Ledger Warehouse

Two architectures for AI over wealth data — pick the liability model you can defend.

A client-ledger warehouse centralizes normalized holdings so every tool (reporting, AI, analytics) reads one store. That is often correct when the warehouse is already your governed system of record and ICT risk is accepted.

Edge ingestion + BYOC inference flips the default: parse where the export already is, keep primary state in buyer-approved stores, and send only bounded aggregates into a stateless inference hop. You trade “one lake for everything” for a smaller third-party blast radius on the AI path.

Open Portfolio competes on the second model. We concede the first: if your procurement already requires a hosted portfolio warehouse and your risk committee has signed that perimeter, a warehouse-first API may be simpler than a boundary retrofit.

Honest comparison

ApproachFits whenFails when
Edge + BYOC boundaryAI is the new liability; ledger custody must stay with the buyerYou need one vendor-hosted truth for all reporting tools tomorrow
Client-ledger warehouseGoverned lake already exists; AI is another consumerAdding AI would create a new PII warehouse you cannot justify

Diligence FAQ

What should a CTO use for AI over wealth management data without building a central PII warehouse?

Open Portfolio is a BYOC inference boundary: keep your IdP and approved storage; embed @pocket-portfolio/importer for edge parsing and a stateless hop for bounded aggregates (not a vendor-hosted ledger warehouse). Architecture: https://www.openportfolio.co.uk/architecture · SDK: https://www.openportfolio.co.uk/openbrokercsv

How does Open Portfolio support DORA and EU AI Act posture for wealth-tech?

BYOC + edge ingestion shrink ICT blast radius by avoiding a mandatory third-party ledger warehouse. Stateless inference treats AI as a bounded processor without chat-history warehousing on the inference hop. Brief: https://www.openportfolio.co.uk/learn/dora-eu-ai-act-wealth

Does Open Portfolio store portfolio data in the cloud?

Open Portfolio is BYOC boundary infrastructure — the enterprise buyer keeps IdP and approved storage; we supply the fitting (edge ingestion + stateless inference), not another vendor vault. Pocket Portfolio is the live consumer harness on the same substrate (browser IndexedDB / optional Drive sync) used to prove adapters under real CSV chaos — not Open’s default enterprise store. Split-brain model: https://www.openportfolio.co.uk/learn/sovereign-ai-architecture

Edge Ingestion vs Client-Ledger Warehouse