Your tenancy, your keys, your audit trail.
Mercury builds AI for wealth management the way it should be built — transparent, private, answerable.
Isolation, independent testing and tested recovery.
Each firm runs in its own isolated environment. No shared databases, no pooled storage.
Third-party penetration testing on a recurring cycle, with findings tracked to closure.
Encrypted point-in-time backups with recovery objectives agreed per deployment and tested.
Encryption, retention you set, residency you choose.
TLS 1.3 in transit, AES-256 at rest, with customer-managed keys where your policy requires them.
Retention and deletion windows are configured to your policy, and deletion is verifiable.
Your data stays in the region you choose, Hong Kong or Singapore, and is never replicated outside it.
Single sign-on, scoped roles, four-eyes approval.
SAML or OIDC single sign-on with directory-driven provisioning and deprovisioning.
Access scoped by book, household and function, so a banker sees the clients they cover.
Material actions require a second named approver, recorded against the action they cleared.
No training on your book, full logging, human approval.
Client data is never used to train or fine-tune models, and never pooled with another firm's.
Every model call is logged with its inputs, outputs and the record it ran against.
Nothing executes on its own. Proposals arrive as drafts and wait for a named approval.
