InCountry for AI
Compliantly operate AI globally with anonymized data and locally with global and local data
Compliant global and local AI
AI has increased the criticality of sensitive data, driving many countries to regulate cross-border data transfers of both AI training and inference data. Medical records are now a predictive roadmap to a patient’s future, as well as personal health information. InCountry enables the compliant use of global AI with anonymized data from multiple countries for training, Retrieval Augmented Generation (RAG), and inference. Local staff can use global AI with permission-based injection of regulated data in responses, as well as tailored local AI containers that combine local and global data.
Operate global AI with anonymized data from multiple countries
- Anonymize regulated data before it crosses borders to train AI models
- Support Retrieval Augmented Generation (RAG) using anonymized data
- Enable local staff to use global AI systems and reinsert PII with response injection while enforcing both fine-grained and coarse-grained permissions
Operate local AI by combining local regulated data with global data
- Use local and sovereign AI models in managed containers that are tailor made for local language, customs, and compliance
- Optionally anonymize data used to train local LLMs
- Detect and redact PII in prompts
- Reinsert PII with response injection while enforcing both fine-grained and coarse-grained permissions
Compatible with Salesforce Einstein AI
- InCountry’s AppExchange Managed Package provides deep integration with Salesforce
- Fully redact regulated data in each country while still maintaining Einstein features like pipeline forecasting
- Recombine global insights with local data in the Salesforce user interface and Salesforce Data Cloud
The turnkey, enterprise-grade solution for data residency
Enterprise-Ready
- Two points-of-presence in each country with active-active failover
- SaaS, single-tenant anywhere, AWS Outposts, and sovereign cloud options
- Guaranteed messaging across unpredictable networks
Developer-Friendly
- Use from Claude Code, OpenCode, and OpenAI Codex
- Flexible tokenization, hashing, masking, and AI generalization algorithms
- Invoke via MCP Tool, MCP Proxy, Web Service Proxy, REST API, or Command Line Interface
Bank-Grade Security
- Granular policy-based authorization combined with fine-grained authorization from source apps
- Secure SDLC and operations with active threat monitoring
- Searchable encryption using NIST standards
Global Compliance
- Data leak prevention across borders and data minimization within borders
- Detailed support for regulatory approval in complex jurisdictions
- Full observability with detailed logging and OpenTelemetry support