Nance uses automated systems, machine learning models, and AI-assisted tooling to support features such as transaction categorization, merchant normalization, spending summaries, suggested insights, comparisons, budgeting assistance, grouping support, and conversational responses through Nancy.
To enable these features, we may process your transaction data, SMS-derived and email-derived financial signals, prompts, typed questions, voice inputs, correction history, budgets, goals, and preferences. This may be necessary to answer your questions, surface personalized insights, trigger supported actions, and improve the consistency of the user experience.
Some outputs are generated probabilistically or through rules plus AI and may be incomplete or inaccurate. We therefore treat many outputs as assistive and user-facing rather than as independently verified official financial records.
Where permitted by law, we may use aggregated, statistical, or de-identified information to improve service quality, model routing, system reliability, and product analytics. Where we use third-party AI API providers, their systems may temporarily process prompts, context, and outputs to provide inference or meet safety and abuse-prevention obligations.