Designing a local-first financial decision workspace that turns fragmented financial data into prioritized, evidence-backed next steps.
Product thesis
Lunera is a local-first personal finance concept for a focused weekly money review: identify what needs attention, inspect the supporting evidence, and test a plan before acting. The dashboard is the entry point, organizing cash position, a priority review, and a safe next action without turning a planning view into automated money movement.
Start with an honest baseline.
The five-step setup captures only the context needed for a first planning brief: identity, a user-entered baseline, optional priorities, wealth context, and a final local review.
Decision loop
The decision loop is deliberately sequential: begin with the Financial Brief, inspect a ranked signal, diagnose the pressure driver, verify its ledger evidence, then review a budget or goal plan. Each module narrows the next question instead of asking the user to act on a single recommendation.
Explainable AI
The AI workspace presents signals as items to inspect: each has a priority, evidence summary, confidence, and a bounded next step. It helps the user review a recommendation instead of treating it as an instruction.
Safe scenarios
Cashflow identifies the main pressure driver before a plan is changed. Scenario and payment sandboxes then let a user test a category adjustment or upcoming payment as planning views, without writing to balances or transactions.
Evidence trail
Transactions provide the traceable record behind a signal, while Budgets translates that evidence into a month-level plan. Each module has a different job: verify the source, then review the plan.
Goal planning
Savings and investments remain planning workspaces, not execution surfaces. The screens expose progress, allocation, and review state so a user can judge the trade-off before committing to a plan.
Reflection
The prototype establishes an information model for a planning-only finance workspace: surface the priority, show the evidence, preview a choice, and retain the review context. It demonstrates the interaction and trust model; it does not claim adoption or financial outcomes.