recurring-transactions

Recurring subscription detection works by grouping Wells Fargo transactions by normalized payee and similar amounts, then inferring the likely frequency and computing a confidence score for each detected pattern.

Yes, you can analyze transaction history over configurable date windows ranging from 3 to 24 months. The skill infers regular payment frequencies and estimates the next expected charge date for each detected payee pattern.

Yes, it requires SerenDB read access to Wells Fargo categorized transaction tables. The detection process writes the resulting recurring patterns, runs, and snapshots into dedicated database tables via a Python entrypoint script.

The best way is to run automated payee-level pattern discovery over categorized bank transactions. This approach detects regular bills, infers their frequency, and persists the recurring patterns into a database for downstream reporting.

Confidence scores are calculated by evaluating the consistency of grouped transactions by normalized payee and similar amounts. The inferred frequency, ranging from weekly to annual, strengthens the confidence score for each detected recurring pattern.