The problem is not adding expenses

Most personal finance apps promise the same thing: connect your accounts, classify your spending, and look at nice charts. The problem is that for many people in the Dominican Republic, financial life does not fit that model.

A person may be paid twice a month, receive remittances, pay informal loans, buy with cash, use cards for specific categories, transfer through mobile banking, and manage part of the household budget over WhatsApp. If an app does not understand that mix, it becomes another place where the user has to work.

AI can help, but only if it reduces friction instead of decorating an incomplete product.

What a local app must understand

Biweekly payroll and irregular flow

Many imported budgeting models assume stable monthly income. In practice, a person may see peaks on the 15th and 30th, extra income from independent work, and heavy expenses early in the month.

The app should project liquidity by windows:

• Today to 7 days.
• Until the next payroll date.
• Until month end.
• Upcoming fixed commitments.

The user does not only need to know that 23% went to food. They need to know whether they comfortably reach the next paycheck.

Cash as part of the system

An app that ignores cash ignores a real part of daily economic life. But asking users to enter every cash expense does not work either.

A better strategy:

• Detect cash withdrawals.
• Ask once for the likely purpose.
• Learn patterns by amount, day, and approximate location.
• Offer fast categories, not long forms.

The goal is not perfect accounting. It is enough clarity to make better decisions.

Remittances and family commitments

In many households, planning is not individual. It includes support for parents, children, school, health, and shared responsibilities. The app should support family goals, shared envelopes, and reminders that do not feel like scolding.

UX matters as much as the model. “You overspent” can create abandonment. “If you move RD$1,500 from leisure to transportation, you cover the week without touching savings” is more useful.

Where AI belongs

Automatic transaction classification

The first utility is simple: convert ugly bank descriptions into useful categories.

POS 004823 SUPERMERCADO NACIONAL

should become:

• Merchant: Supermercado Nacional
• Category: Groceries
• Type: likely recurring consumption
• Confidence: high

When confidence is low, the app asks. When the user corrects it, the system learns.

Change detection

A good app does not only show a budget. It detects shifts:

• Delivery spending increased for three weeks.
• A new recurring charge appeared.
• A telecom bill looks duplicated.
• Balance is dropping faster than in previous cycles.

That is more valuable than a pie chart.

Recommendations with limits

AI can suggest, but it should not manipulate. Financial recommendations should be:

• Explainable.
• Reversible.
• Based on recent data.
• Clear about uncertainty.
• Not selling debt as the first answer.

An app that pushes credit whenever it detects stress will lose trust quickly.

How we would build it at SSD

The minimum architecture has five pieces:

1. Ingestion of bank, POS, or manual data.
2. Transaction normalization.
3. Category and pattern engine.
4. Recommendation model.
5. Consent, privacy, and audit layer.

The product should work even before full Open Banking is available. It can start with statement imports, private integrations, or PuntoOS data for business owners. When open APIs mature, the experience is already ready.

The right KPI

We would not measure success by “users who connected an account.” That is technical activation.

We would measure:

• Users who understand projected balance.
• Users who reduce forgotten charges.
• Users who complete a savings goal.
• Users who detect a problem before overdraft.
• Users who return without aggressive notifications.

The winning Dominican finance app will not be the one with the most visible AI. It will be the one that makes people feel less uncertainty when they open their bank.

Context sources

• Superintendencia de Bancos: bank account growth in the DR
• Banco Central: financial inclusion