Software AG: keep up with big fast data - Richard Bentley
In an industry where customer service is becoming the most important differentiator, standing out from the crowd requires a bank to make the right decision in as close to real time as possible. Whether for improving customer service or enhancing fraud prevention processes, Richard Bentley, VP financial services at Software AG, explains why banks have to get to grips with not only the amount of data they hold, but also the speed at which it flows.
Everyone in the financial services industry has become familiar with the words 'big data', though the term has many different definitions. The big picture is that banks have a wealth of data at their fingertips, but are not yet wringing out all of its potential value. When encouraged to grasp the nettle of big data, banks are often told of the scale of the in-memory information repositories at their disposal, but what is often overlooked is that the data is not static - it comes in rapidly flowing streams.
Data in huge volumes, travelling fast, may seem overwhelming, but software vendors have been working hard to help banks tackle 'big fast data' (BFD) with the capacity to perform complex event processing (CEP).
"Software AG talks about BFD, which is a confluence of messaging, visualisation, in-memory and CEP, and it was quick to realise that getting value out of it was about dealing with the velocity and veracity of that information," says Richard Bentley, VP financial services at Software AG. "It was already a leader in in-memory data management and universal messaging, but recent acquisitions have brought in the capacity to deal with real-time and contextual data.
"Software AG has filled in the missing link in its portfolio, so it can create business solutions that deliver value in areas like fraud management and customer experience, which are BFD problems. The value of data in these areas can depend on analysing it in milliseconds or even microseconds.
Solving the Goldilocks problem
Software AG's purchase of Apama, which specialises in CEP, gives it the capacity to help financial institutions respond with timely and relevant actions to operational and transactional events, including suspected instances of fraud. Apama's event processing platform serves as a complete design and deployment environment for CEP applications, providing analysts, developers and administrators with a full set of real-time analytics tools to improve processes such as fraud prevention and assist with the development of specific solutions for algo trading, forex, e-commerce and market surveillance.
"Apama came out of research at the University of Cambridge into the processing of high volumes of streaming data in order to find patterns, including time-based patterns, that could prompt immediate action. It is one of the pioneers of CEP and has matured over 15 years to the stage where it has provided applications to tier-1 banks in areas such as electronic trading and, now, real-time surveillance, risk management and fraud detection," explains Bentley.
"Software AG realised that the interesting problems to solve were all bringing real-time decision-making capability and business rules to data sources, including the data in legacy systems. Fraud management was a priority from the start as it is an area where automisation really pushes the boundaries of technology and drives innovation."
Fraud management is a challenging issue for many reasons. To begin with, it demands rapid analysis of and response to large volumes of streaming data. Then, once that capability is in place, it requires the parameters for automated decision-making to be set at the right level to respond to suspected fraud appropriately without impeding a customer's interaction with the bank.
"Fraud is a Goldilocks problem," says Bentley. "It is hard for financial institutions to get fraud management right because if they do too little then they lose money to fraudsters, but if they do too much they will suffer customer attrition because people will have their cards blocked too easily, for example. Banks have to understand customer activity in real time and in context in order to respond within a shrinking intervention window."
There are many questions that must be asked of fast-flowing data in order to spot fraud. For instance, is it surprising that a customer has withdrawn the maximum daily allowance from an ATM? Where is that ATM and has that customer used it before?
To answer such questions it is crucial to take into consideration not only the real-time event, but also the historical context for that individual customer.
"The solution can also look at external data feeds as well as looking internally at existing data within the bank," notes Theo Priestley, vice-president and chief evangelist at Software AG. "For instance, it can look at social network channels or mobile phone accounts and GPS data for information that might show that a customer is in a different country to that in which the ATM is being used.
"You will never get fraud down to zero, so you have to ask what an acceptable level of fraud is," he adds. "What is certain is that fraud is on the rise in many areas of financial services, particularly as services move online. We have to focus on new types of fraud as well as the traditional ones. The problem is always accretive - the number of methods for perpetrating fraud never reduces."
In short, the financial services industry has to hit a rapidly moving target in a high crosswind that is always changing direction. Software AG, however, believes that it has the technology to let banks do just that.
Understanding problem and solution
One of the biggest challenges that Software AG understands it must deal with is the shift towards omnichannel banking. Customers expect rapid and consistent service regardless of which channel they choose for interaction with their bank.
"Banks have to be reactive to threats of fraud and must respond in an appropriate timescale," says Bentley. "The shift to an omnichannel environment has forced a reconsideration of fraud management, which must be managed in a layer that sits across all activity, not just credit cards, for example. Software AG's advantage in this area is its agility and its ability to evolve in line with changing regulations and changing customer expectations.
"Banks have to be compliant, but they must balance this with the need to be responsive and innovative," he continues. "There is a lot of demand for our solution. Banks don't want point solutions that are applied to just one channel. Attacks from fraudsters can happen across multiple channels, so we have made a platform that sits across all channels and, importantly, continually evolves as the threats change."
Another key advantage of working with Software AG in this space is the speed with which it can implement its solution. At DBS Bank, the largest bank is Singapore, it delivered a real-time solution in just six weeks following high-profile incidents of fraud. For that project, DBS won the best security and risk project from The Asian Banker, and is now rolling out the solution to other regions in Asia.
"We know about responsive and scale; this makes us a leader in the BFD area," remarks Bentley. "We are known to be able to solve these problems for our customers and reputation is everything in this industry. Furthermore, we don't favour a 'rip-and-replace' approach, which is important in addressing the omnichannel environment. If a bank had to replace its core systems for each channel, it would be a nightmare. We keep the expense down by making our solutions simpler to deploy.
"We layer our solutions on top to extract the right data from existing systems. This non-invasive, accretive approach is important, as it means our solutions complement what our customers already have. If a bank's fraud management systems are not adequate, it will suffer financial loss and damage to its reputation, but we can turn that around, as we did with DBS."