Back to Directory
MODERATE RISK

Salesforce for Housing

Global (customer-region selection) ·Social Housing
OVERALL RISK SCORE 68/100
SOC 2Compliant
GDPRCompliant
Trains on User DataYES — HIGH RISK
Enterprise TierAvailable
Data Retention365 Days
Data ResidencyGlobal (customer-region selection)

Risk Summary

Salesforce's Housing Cloud and Nonprofit Cloud are deployed by several large UK housing associations for tenant relationship management. Salesforce's Einstein AI layer can process tenant interaction history, case notes, and service requests stored in the CRM. The principal concern is that Salesforce's AI models are trained on aggregated data across customers — even with Enterprise Shield, the default model improvement programme may expose patterns derived from tenant data. Einstein GPT features add generative AI risk on top of existing CRM data stores.

Technical Briefing

Salesforce Einstein processes tenant case data, communication history, and service request patterns for AI-driven insights. The Einstein Trust Layer claims to mask PII before sending data to LLM providers, but the masking effectiveness for unstructured housing case notes is unverified. Enterprise Shield offers tenant-specific encryption keys and restricted model training, but requires explicit configuration — not default. 365-day data retention.