≡ Menu

IBM Big Data & Analytics: Analytics and Fighting Fraud

Share

An update next on IBM’s solution for counter-fraud.  For IBM fraud involves an intentional act that misrepresents and is both illegal and designed for financial gain. They further divide this into organized and opportunistic fraudsters for whom different kind of detection is required. Fraud is growing problem because:

  • Fraud schemes are increasingly sophisticated with organized crime moving it it due to both opportunity and lower risk of violence/arrest
  • Fraud is no longer acceptable as a cost of doing business – regulatory environment is tigher for instance
  • Consumers expect less fraud and are increasingly cranky about it, reporting it more readily for instance.

Schemes are getting sophisticated involved hacks, networks of people globally and laundering all being used together such as one that changed ATM limits on some compromised accounts, used people in many countries to withdraw and launder millions of dollars.

Attacking fraud can require the classic decision management OODA loop – Observe, Orient, Decide and Act – where fighting fraud requires tightening your OODA loop to be faster than the fraudsters’:

  • Observe increasingly requires big data to broaden the range of signals being observed
  • Orient requires increasingly advanced analytics
  • Decide is all about Decision Management, embedding analytics into decisions
  • Act is about connecting to the business processes and case management

Counter fraud is one of IBM’s signature solutions this year. While Fraud is connected to risk and security, overlapping in operational risk and cyber security, fraud is a separate problem also. This requires elevating the agenda for fraud detection and prevention to the company level so it becomes proactive not reactive. It requires an ability to act with insight and to adapt with agility – you have to have a tighter OODA loop to stay ahead.

It also requires four capabilities:

  • Detect in time to take action
  • Respond with a plan and supporting technology when you detect something
  • Discover patterns of fraud from historical analysis
  • Investigate suspicious activities to confirm and resolve issues

IBM has developed a framework for fraud that handles prevention and countering fraud. Prevention involves both IT security and business controls. Countering includes detection, response, investigation and discovery. All of this wrapped with intelligence (analytics) and governance. This framework involves a number of acquisitions and product development efforts and focuses them all on counter fraud.

IBM has made this counter-fraud stack available has a single product offering with value-based pricing for Financial Crimes in Banking, Insurance, Healthcare and government. These products are supporting with specific partners and consulting resources. These solutions contain multiple analytic techniques from entity analysis and link analysis to geospatial and behavior analytics, predictive analytics and decision management. Once a potentially fraudulent transaction is identified by these various kinds of analytics it can be pushed into the case management environment. Case workers can access the various analytics, interacting with them to follow links found by the link analytics for instance or using text analytics to work through large amounts of text associated with the case.

IBM has also been investing in helping clients get better at detecting fraud with new services offerings (accelerators, target operating models and scaling services) as well its new Red Cell team to act as a central hub to disseminate information about fraud rings and other challenges.

Share

Comments on this entry are closed.