The firm's new compliance platform uses predictive IT to reduce false alerts and help compliance teams focus on activities that pose the highest risk, while increasing something it calls 'security coverage.'

The latest platform includes the Risk Meter, apparently the first piece of social media compliance software that uses machine learning to prioritise "highest-risk offences" for supervision teams. Machine learning allows Hearsay to take the nearly 100 million interactions it processes every year, from more than 150,000 advisors and agents, and use that intelligence to predict the highest-risk items.

Hearsay’s compliance platform seeks to fulfil the requirements of regulators such as those from FINRA, the SEC and Federal Financial Institutions Examination Council. It also handles each firm’s policies to help it track and supervise advisors’ digital activity and communication. It looks at advisors' interactions with customers on social media, text and email.

Hearsay’s Risk Meter is the first piece of 'social compliance' software that uses machine learning to prioritise violation alerts, allowing teams to deal with the most risky items first. Additionally, its Intelligent Alert system can spot violations using not only prohibited words, but also other data such as the possibility that similar activities in the past were considered to be true alerts; this ought to reduce the number of false alerts. The software decides on the things that are most important for the organisation to monitor and intercept.

It also does the following.

• Auto-remediation. Hearsay can automatically delete activities performed by an advisor that are clearly against the firm's policy. For example, if an advisor publishes a Facebook post natively through Facebook, thereby breaking company rules, the Hearsay platform can delete that post automatically.
• Smart lexicon. This feature uses machine learning to help firms set rules more accurately and reduce the number of false alerts. For example, if an advisor publishes a post with the text, “I guarantee tomorrow’s barbecue will be amazing,” and his firm’s policy prohibits use of the term “guarantee,” Hearsay’s algorithms can analyse the post’s full context and other data before deciding whether or not to generate an alert.
• Flexible sampling. This is to do with 'spot checks' on content that lexicon-based monitoring did not identify as suspicious and is a way of ensuring that risks do not "slip through the cracks." Hearsay says that one-third of its customers use its content-sampling software as an additional compliance checkpoint.