Post by @ayatalis

Banks use linguistic filters to automatically flag transactions for review. ​Not just amounts. Words. ​This is compliance infrastructure – built to satisfy anti-money-laundering regulations. The filters run automatically. No human reviews the transfer until the flag is already raised. ​Three categories that trigger them most consistently: ​Cash references in digital transfers. Writing “cash back” or “cash for” in a transfer note when no physical cash is involved creates an immediate inconsistency flag. Digital transaction described in cash terms – the system flags it. ​Loan language between private individuals. “Loan repayment,” “paying back,” “money I owe you” – these phrases between non-business accounts flag potential undeclared lending. Tax authorities in multiple countries treat informal loans as potential income if undocumented. The note alone doesn’t prove anything. But it opens a file. ​Vague service descriptions. “Services rendered,” “consulting,” “work done” – without a corresponding business registration on either end – pattern-match to known structuring behaviors. The vaguer the description for a significant amount, the higher the flag probability. ​What the investigator recommends instead: ​For personal transfers – use specific, human language. “Birthday gift.” “Splitting dinner from Saturday.” “Half of the Airbnb.” Mundane and specific. Filters catch evasion – not ordinary human behavior. ​For business transfers – match the note exactly to an invoice. One line. Invoice number if possible. ​The note field feels irrelevant. ​It isn’t. ​Three words in that box can open a review that takes six months to close. ​Write carefully. #information