Making efficiency the new normal: end-to-end automation in financial services

This is a guest blogpost by James Loft, COO, Rainbird.

If 2020 has been anything so far, ‘unpredictable’ is as good a descriptor as any. The financial services sector has suffered disrupted supply chains, rising unemployment, and increased government intervention. Operational resilience is being tested more than ever before. In fact, just last month, the Financial Conduct Authority announced it is conducting a survey on firms’ resilience post-Covid-19.

The need for connected, resilient systems that can accurately manage data from multiple sources is only becoming more important. It’s not surprising that in an EY survey of businesses preparing for a post-Covid world, 41% said they were investigating accelerating automation. The long-term response of the financial services sector to the pandemic will be pivotal in the lives of many, so it needs to be strong.

What is driving the need?

Fraud investigation processes, for example, are still highly manual—in particular, within large banks and insurance businesses. Operational costs are high, due to an unpredictable volume of cases, and increasing complexity means that the speed and quality of responses often falls below necessary standards. Over 90% of cases flagged as fraud will be false positives, which require individual review by large teams.

The Covid-19 outbreak is only likely to exacerbate these problems. According to a Wall Street Journal report, the overall number of attempted fraudulent digital transactions in the US rose 35% in April 2020 (in terms of dollar value), compared with the same month in 2019. Our reliance on online transactions as a result of global lockdown, coupled with the huge numbers of people already experiencing financial struggles, means there is a doubly urgent need for the systems we use to be advanced enough to identify and address such cases.

However, automation in the financial sector has been historically unreliable due to a lack of standardisation and poor data quality, along with the ever-increasing complexity of assets. In fact, in a 2019 survey by ABBYY, 28% of UK businesses not yet using robotic process automation (RPA) said ‘they wouldn’t know what to do with it.’

Use case: fraud investigation processes

End-to-end automation can make fraud investigation 100 times faster and 25% more accurate. Robotic process automation (RPA) can effectively take over manual collation of case data from company systems, before passing it to an automated decision-making platform.

Rainbird has developed with Blue Prism an automated fraud system that replicates the end-to-end fraud investigation process within major organisations. This introduces unprecedented speed, accuracy and scalability, and full auditability (with an explanation for every conclusion). False positives—the bane of so many AI and automated fraud systems—are also greatly mitigated.

For example, RPA bots can refer to a decision made by Rainbird’s intelligent automation platform, and act on the answer without intervention by a member of the fraud investigation team. This can result in a 60% reduction of back office costs (crucial savings at times of economic uncertainty). This process also removes major bottlenecks within complex transactional processes.

In essence, wherever employees are making decisions that are commercial, transactional and high-volume, businesses can use an integrated solution to augment the capacity of human specialists. With full audit trails and explainability, process automation is no longer something to be feared but embraced.

There is clearly increased strain on people completing these manual investigations. End-to-end automation can model and scale the best-practice expertise of fraud investigators, easing the pressure.

The tide is turning

Tellingly, a report by Gartner has predicted that by 2024, organisations will lower operational costs by 30%, through a combination of hyper-automation technologies and redesigned operational processes.

With that said, integrating this kind of end-to-end automation shouldn’t be about saving resources for the sake of it – it should be centred on the adoption of practical, efficient tools that improve business processes as a whole.

It’s ultimately humans who stand to benefit from the advent of modern technology, provided it is implemented fairly and transparently, with the expertise of subject matter specialists at its core.

Data Center
Data Management