Different documents.
Different threat model.
RSP protects operational workflows (complaints, customer correspondence, internal processing). But legal agreements, board papers, and investor materials carry a different risk: contextual fingerprints that allow re-identification by sophisticated counterparties. HCPD extends RSP to handle these high-sensitivity documents.
High-sensitivity documents carry contextual fingerprints
Legal agreements, board memos, and investment materials contain domain-specific details that a sophisticated counterparty could use for re-identification, even after entity-level protection.
Entity-level PII is fully protected by RSP. However, this document type (a legal agreement shared with external counterparties) contains domain terminology and organisational details that a sophisticated reader could use to narrow identification. HCPD addresses this specific threat model.
See HCPD in action
Select a sample document to see how HCPD detects contextual signals that entity-level detection misses.
The right protection for the document type
The regulatory basis for contextual protection
Entity-level PII protection is no longer sufficient under modern privacy regulations. Regulators now require you to address re-identification risk through contextual data combinations. Here is the regulatory basis.
Personal data means any information relating to an identified or identifiable natural person. A person is identifiable if they can be identified, directly or indirectly, by reference to an identifier or to factors specific to their identity.
To determine whether a natural person is identifiable, account should be taken of all the means reasonably likely to be used, including combination of data points that, taken together, allow identification.
The ICO applies a "motivated intruder" test: could a person without specialist knowledge but with access to resources such as the internet, public records, and social media re-identify individuals from the data? Contextual fingerprints fail this test.
The CCPA/CPRA defines personal information to include data that is "reasonably capable of being associated with, or could reasonably be linked, directly or indirectly, with a particular consumer or household." Contextual signals create this linkability.
The Gramm-Leach-Bliley Act requires financial institutions to protect nonpublic personal information (NPI). Contextual data that permits re-identification of account holders falls within NPI scope.
Virginia (VCDPA), Colorado (CPA), Connecticut (CTDPA), and 12 additional state privacy laws all adopt variations of the "reasonably linkable" standard. The trend is towards broader interpretation of identifiability.
The Federal Reserve's SR 11-7 requires banks to understand and manage model risk, including the risk of data exposure through AI model inputs. Contextual re-identification represents an unmanaged model risk.
Banks using third-party AI models must assess data exposure risk. Sending contextually identifiable data to third-party LLMs creates a third-party data risk that must be mitigated.
The PRA expects firms to identify and manage model risk including data leakage through model inputs. HCPD addresses the contextual component of this risk.
The FCA's 2025 guidance on AI in financial services explicitly references the risk of indirect identification through contextual data combinations when using third-party AI services.
| Risk Vector | Entity-Level Only | Entity + HCPD |
|---|---|---|
| Direct identifiers (name, SSN) | PROTECTED | PROTECTED |
| Quasi-identifiers (DOB, postcode) | PROTECTED | PROTECTED |
| Domain terminology | EXPOSED | NEUTRALISED |
| Jurisdictional signals | EXPOSED | NEUTRALISED |
| Temporal markers | EXPOSED | NEUTRALISED |
| Organisational fingerprints | EXPOSED | NEUTRALISED |
| Combinatorial re-identification | HIGH RISK | MITIGATED |
| GDPR motivated intruder test | FAILS | PASSES |
Operational documents. Legal documents.
One platform. Full coverage.
RSP for day-to-day processing. HCPD when the threat model demands it.