Where AI may assist
- Summarizing applications, reports, public pages, and operational records.
- Extracting draft event, volunteer, form, or program content from customer-provided sources.
- Suggesting classifications, narrative drafts, insights, or next steps.
- Enriching public nonprofit profiles from public sources when enabled.
Human decision authority
People remain responsible for eligibility, review, funding, award, payment, employment, compliance, and other consequential decisions. AI output should be checked against source records and corrected before use. FoundationOS does not represent generated content as independently verified fact.
Data handling
Only content needed for an enabled AI request should be sent to an approved provider. Fundamentl does not intentionally use customer content to train FoundationOS models or authorize providers to train public models on customer content. Provider processing and retention remain subject to the configured service and contractual terms.
Customer controls and responsibilities
Customers should limit sensitive data submitted to AI-enabled workflows, review output before relying on it, and tell Fundamentl about contractual or regulatory restrictions during onboarding. AI availability and provider choices may change as controls and product capabilities evolve.
Known limitations
- AI can omit context, misclassify information, repeat source errors, or generate incorrect statements.
- A summary may not preserve every qualification, exception, or attachment.
- Public-source enrichment can be stale or conflict with organization-submitted information.
- AI-generated recommendations should never be the sole basis for a consequential decision.
Questions
Contact privacy@fundamentl.app for data-processing questions and support@fundamentl.app for workflow-specific AI questions. Material changes to AI data handling will be reflected in this disclosure.