Hewlett Packard Enterprise (HPE) released a product this week designed to help financial institutions minimize the risk of regulatory fines. The software, Investigative Analytics, aims to help keep heavily regulated financial firms compliant using machine learning to understand what constitutes risky and non-risky behaviour.
Investigative Analytics, which runs either as a SaaS service or as a managed service in the customer’s own data centre, will analyze information from different data sources, including email, instant messaging, and trading systems, according to the firm. HPE noted the solution will also analyze voice archives.
The system uses machine learning to create a baseline showing normal, ‘clean’ patterns in regulated activities such as trading. It can then spot anomalies by measing new activity against a set of “key risk indicators.”
These KRIs include things like the social distance between parties in a trade (you might be suspicious if a trader kept doing the same kind of deal with someone he went to school with, for example), and the timings of communication (i.e. “Why is one trader always conducting a particular kind of trade outside office hours?”)
“HPE has had a high degree of success in identifying issues which were previously unknown using current communication supervision techniques: current results show this approach to be up to 200 times more successful in catching true issues,” said Joe Garber, vice-president of marketing in Information Management & Governance at the firm.
That’s all well and good, but baselining typically requires a base set of data that you know to be clean. What happens if the trading data that a customer uses for baselining is stuffed with rogue trades and specious practices?
“Baseline behavior comes from reviewed sets of material which have been assessed as part of the regulatory compliance workflows which have been in operation for over a decade,” said Garber. “High confidence clean sets — based on historic audit records — are used to establish a confident baseline.”
Other companies have been getting in on the market for automatically detecting financial anomalies, too. In December, Software AG released the latest version of its Apama Capital Markets Foundation toolset. Unlike HPE’s offering, Apama is a developer platform that allows financial institutions to build their own applications, and simply folds in the compliance analytics.
It’s no surprise that software vendors and service providers are clamouring for a piece of this action. The market for financial compliance is healthy. A 2014 Everest Group report on IT outsourcing in capital markets found that 89 per cent of the deals had a regulatory compliance component, and 82 per cent of them had an analytics element.
HPE is pricing Investigative Analytics based on the amount of analytic compute power its customers use on a monthly basis.