AI Transparency Notice
How Indavera uses AI in financial crime compliance workflows — last updated: [date], DRAFT, not yet in effect
Indavera uses large language models for specific assistance functions within the platform. Current functions include assisted transaction-column mapping, investigation analysis, evidence verification, regulatory narrative drafting, pre-approval quality review, rule-authoring assistance, policy question-answering, case summarization, onboarding document extraction, information-request drafting, onboarding recommendations, and compliance-document drafting.
Human oversight
AI outputs are advisory. Indavera is designed so that AI does not silently replace deterministic calculations, configured rules, permissions, report validation, or required human approvals. Customers remain responsible for decisions made using the platform.
Evidence and verification
Where supported by the workflow, Indavera links material findings to underlying evidence and uses a separate verification step to identify unsupported or contradicted investigation statements before human review.
Data sent to AI providers
The platform currently integrates Anthropic and OpenAI. Only information relevant to the invoked feature is intended to be sent, but this may include customer/business data, transactions, case information, policy content, or uploaded onboarding documents.
Provider data handling
Indavera currently uses stateless AI calls and does not use persistent-memory or provider file-storage features for the documented AI workflows. The platform includes controls intended to request restricted provider use and data-region handling. Provider-side request logs, retention periods, and training restrictions remain governed by Indavera's applicable contracts and provider terms; no stronger zero-retention or no-training guarantee is stated unless contractually verified.
Limitations
AI can make mistakes. It may misunderstand documents, omit relevant facts, or generate incomplete language. Users must review AI-assisted outputs before relying on them for compliance, onboarding, risk, or regulatory decisions.
Questions about AI processing: support@indavera.com