AI Transparency Statement

Where AI is used

foretale.ai uses AI to assist with document analysis, risk and compliance workflows, natural language queries over authorized data, summarization, and test recommendations.

Model providers

Primary inference runs on Amazon Bedrock with organization-configurable models. We do not use customer content to train third-party foundation models. Optional observability tooling (e.g., LangSmith) may process prompts and outputs when explicitly enabled.

Data used by AI features

AI features access only data authorized for the authenticated user and project. Context is scoped per session and organization. Attachments and query results are processed transiently and stored according to project retention settings.

Safety and guardrails

Amazon Bedrock Guardrails and application controls filter harmful content and constrain off-topic responses. Rate limits protect against abuse.

Known limitations

AI outputs may be incomplete, outdated, or incorrect. They should not be treated as legal, audit, or regulatory advice. Users must verify outputs against source records and apply professional judgment.

• Models may hallucinate facts not present in source data.
• Complex regulatory interpretations require expert human review.
• AI does not replace formal audit procedures or sign-off requirements.

Contact

Questions about AI use in foretale.ai: contact@hexango.com.