Swift, the interbank messaging system that hackers used to steal $81 million from Bangladesh last year, is developing a payment-screening service that will allow small member banks to automate the flagging of suspicious payments.

The new service is part of a series of measures from the Society for Worldwide Interbank Financial Telecommunication to defend against cyberattacks that aim to fraudulently use banks' connections to the messaging system. It will place a “red flag” on payment messages that appear risky and spot anomalies, provide real-time alerts and allow customers to put a hold on unusual messages, the cooperative said in an emailed statement on Tuesday.

Swift aims to begin offering the service by early 2018. Its price will depend partly on how many banks adopt it, the cooperative said.