Some in the finance and banking industry are kicking goals with robotics process automation (RPA), AI and business process automation, but the industry still has few use cases in production, according to a recent report from Capgemini, “Growth in the Machine: How financial services can move intelligent automation from a cost play to a growth strategy.”

The report is based on a survey of more than 1500 financial services executives globally, including the US, UK, India and Europe. The sectors it focused on were retail and commercial banks, capital markets – and life and non-life insurance. Of the organisations, 42% had global revenues greater than $US10 billion.

It concludes that “While organisations have driven significant efficiency gains through RPA … many struggle to make a success of intelligent automation, failing to achieve scale and drive value.”

Of the organisations surveyed just 10% have implemented automation at scale, i.e., across all the geographies and processes that the company operates in. While 17% were stuck at pilot stage, a large number (69%) have managed to get past pilot stage to deploy at one or two sites, but large scale adoption across business processes, functions, or geographies is still elusive

Looking across all stages of implementation – from concept to full-scale implementation – Capgemini found that 40% have adopted RPA, but only 4% have adopted machine learning.

It found India leads in deployment of intelligent-automation initiatives at scale.

This reflects significant investments and the availability of talent.

“It’s not surprising that India leads in automation deployment,” says a senior executive from an Indian lending, wealth management, and insurance firm.

“The automation technology capabilities can be built within a short span of time, but domain expertise is the differentiator for success. Talent is definitely not an issue for automation in Indian financial services.

Malaysia-based CIMB Bank started RPA implementation in September 2017 for automation of a host of banking operations, including financial reconciliation, maintenance, e-banking, audit confirmation, and auto checking/reminder follow-ups. To date, it has already automated 47 processes, with a marked improvement in the processing times. Additionally, the bank has managed to achieve an impressive 90% reduction in turnaround time for nine out of 15 banking processes that were automated using robotic process automation (RPA) in November 2017.

Around half of organisations surveyed face issues in integrating their automation platforms with legacy software systems and tools. For an automation program to run successfully, it needs to have access to all the source data from multiple systems. This is particularly true for those processes, such as payments, that rely on access to data from multiple systems. Getting and maintaining that access is complicated by many factors: technology integration, change control, and other risk and control issues.

“We have had a lot of technical challenges with different platforms and discovered that some platforms are not very robot friendly,” says Jenny Dahlström, head of Robotic Implementation for Handelsbanken Capital Markets, a boutique investment banking firm.

“We faced problems with traceability, access rights, and firewalls.”

Inability to establish a business case

Capgemini found that 43% of organisations say they do not have a process for identifying the business case for automation. This reflects the lack of focus on high-impact use cases. When teams execute complex use cases with long payback periods, senior leaders’ backing for future pilots is challenged. In fact, 41% say they are struggling to get leadership commitment for advanced automation.

Gaining consensus across the business – on issues such as the processes to be optimised or clarity on roles – has been particularly challenging for nearly half of organisations.

Units are often reluctant to relinquish autonomy in areas they consider to be critical to their existence.

“Split ownership of processes is a concern, especially when front-end or online processes are not integrated with other processes,” says Nils Henrikson, program director Digitization and Claims Automation for Trygg-Hansa, a Scandinavian insurance company.

“It is important to have end-to-end ownership of a process. It means firms can drive automation initiatives with full control and better coordination.”

Algorithms require the right data at sufficient volumes

Although industry players have made significant investments in data governance programs focused on quality and availability, Capgemini found that 46% said that the lack of an adequate data management strategy hampers progress

“We have huge amounts of data in our business, and we are not really using it to a full extent,” said Trygg-Hansa’s Nils Henrikson. “I think we can do a lot more with better use of data with machine learning.”

Around half of respondents said cybersecurity and data privacy are major factors preventing action. Privacy concern stems from the potential misuse of data containing customers’ personally identifiable information (PII), such as names and addresses.

Capgemini estimates that Intelligent automation could add $US512 billion to the global revenues of financial services firms by 2020, with $US243 billion gained in insurance and $US269 billion in banking and capital markets.

The full report is available HERE