Robots take over retail banks’ back offices

• Robotic process automation refers to logic-driven robots that execute pre-programmed rules on structured and some unstructured data
• Banks benefit the most when they focus on automating activities with higher volumes
• Implementing RPA sooner rather than later, and across as many processes as possible, is essential for competing effectively in an increasingly dynamic retail banking environment

In March 2016, the Royal Bank of Scotland laid off more than 500 staff in its wealth management division and replaced them with robo-advisors. This year, Fukoku Mutual Life Insurance in Japan is one of many financial institutions laying off staff, with more than 30 staff being replaced with an artificial intelligence system. The company expects to increase productivity by 30% while saving about $1.2 million (¥140 million) per year.

Robotic process automation is software-driven
Robotic process automation (RPA) refers to logic-driven robots that execute pre-programmed rules on structured and some unstructured data, as PWC describes it. The focus for RPA is to automate mundane, repetitive, and manually-intensive processes so that banks can improve efficiency, increase accuracy and operate 24/7. At the highest end, robots could eventually learn from prior decisions and data patterns so they can make decisions by themselves.

It is important to understand that these robots are not the cute consumer-focused robots people often envisage, EY Singapore partner Liew Nam Soon told The Asian Banker: “This is software algorithms, a pieceof software that does manual activities really well. What the software does, as long as you provide the data, is the manual key-in, plus checking, plus reconciliation. It’s complementary to process workflow.”

The impact of RPA
Retail banks that have implemented RPA have been getting impressive results and a quick payback. Banks are already getting robots to operate at a fraction of the cost of an offshore worker and achieving 25% to 40% cost savings net of software costs said Liew, or even greater cost decreases in highcost location such as Hong Kong. Moreover, there can be a 100% increase in quality, and RPA also frees up existing people for more value-added work, decisionmaking and analysis. Accenture found even greater benefits in some cases, with RPA in financial services reducing handling time in processes by 40% while decreasing processing costs by up to 80%.

PWC said the benefits also include enabling staff to shift their focus to innovation or strategy and other business development activities, providing shorter cycle times and faster delivery than human workers, increasing quality that automatically improves the customer experience, and providing 100% accuracy.

They calculated that a robot’s efficiency is equivalent to the throughput of approximately 2 to 2.5 full-time equivalent (FTE) staff in the financial services industry because it can complete a task in a shorter time than a person.

The impact on headcount can be huge. Deloitte said one large bank it worked with redesigned its claims process and deployed 85 software robots to run 13 processes, handling 1.5 million requests per year. The bank was able to add capacity equivalent to about 230 full-time employees at approximately 30% of the cost of recruiting more staff, and it recorded a 27% increase in tasks performed correctly the first time.

Looking further ahead, Citi forecasts that much of the cost of branches and associated staff costs, which make up about 65% of the total retail cost base of a larger bank, can be removed via automation. While the pace of staff reduction through 2015 had been gradual at about 2% per year and had resulted in an 11% - 13% reduction in staff compared to levels before the financial crisis, Citi believes that the pace of headcount reductions could increase to 3% per year and result in an additional 30% reduction in staff by 2025. The total reduction in staff between about 2005 and 2025 could
then be as high as 45%.

The benefits also go beyond just headcount. At ICICI, for instance, software robots are now processing more than one million transactions per day. While the ways customers interact with the bank have not changed significantly after the introduction of software robotics, ICICI senior general manager, Anita Pai, shared that customers have seen the bank’s response time improved by up to 60% and accuracy has increased by as much as 100%.

Given the benefits, the pace of implementation is increasing and spending on RPA may well rise more than ten-fold over the next five years.

Where to implement RPA
While RPA can be used across many functions in the back office in retail banking and even in some customer-facing functions, banks benefit the most when they focus on automating activities with higher volumes.

PWC said the ideal candidate for digitisation is work that is repetitive and logic-driven or rules-based, with RPA used to bridge between various legacy systems where work has been done manually. These processes can include management reporting, client reporting, account opening and reconciliation. RPA can also be used in processes where the bank needs to scale up quickly when extra support is needed and scale down quickly when volume decreases, as happens around onboarding customers or processing transactions during marketing campaigns or the launch of new services.

What doesn’t work as well, PWC has found, is automating processes that are undergoing change, such as a technology application that’s set to undergo a significant change. Some activities may also be too technically complex or costly to automate without other  foundational changes such as data standardisation.

In Southeast Asia, EY Partner Liew Nam Soon told The Asian Banker, banks are gravitating towards RPA in certain functions - finance, payments, and the front office. While robotics was thought of as for the mundane repetitive manual tasks in the back office, RPA is being increasingly applied for client onboarding, KYC, and AML in the front office. Although the early movers were multinational banks, Liew said regional and domestic banks have caught up. Even in countries such as Malaysia and Indonesia, where it has been easier to “throw people at the problem,” companies are looking at robotics. “Hiring people is tough. RPA can bridge the gap.”

The focus, then, is on selecting functions with less complexity and volatility that also have large numbers of staff performing mundane functions. It is also important not to automate simply for the sake of automating, Cognizant’s president of banking and financial services, Prasad Chintamaneni noted: “Banks should start by reimagining the entire business process. Once that is done, robotics can take it to a whole different level.”

When banks use RPA to automate the customer onboarding process, for example, they can boost customer satisfaction, increase retention and drive greater share of wallet. Processes that could benefit the most from RPA include mortgage processing, claims processing and customer service. ICICI Bank, for instance, started with low-complexity processes in retail operations, then moved on to medium-complexity, followed by automating highly-complex processes. “We looked at volume, repetitiveness, and could one application interact with an external application,” Pai said.

Best practices for implementing RPA
When done well, RPA can be implemented quite quickly. While it can take six to twelve months to automate some processes, ICICI said its extensive experience has made the process even faster, with the bank now being able to set up some new processes in a week or 10 days.

What banks need to do to maximise the benefit is to start by looking at the big picture, understanding the fundamental data and workflows so they can figure out which processes have lower efficiency, and identifying where the actual challenges are. It is also essential to make sure there’s a roadmap from proof-of-concept (POC) to production. Whilst it’s easy to prove robots work and deploy them for an entire finance or operations department with hundreds or thousands of staff, Liew said it needs to be thought out properly and there needs to be clarity about the ultimate end game in terms of deployment.

While robots can be used to fix specific problems, PWC said its experience shows that banks will see the greatest benefits by thinking about deployment at the enterprise level and prioritising programs on a bankwide basis. Focusing on stable high-volume processes that are rules-based and do not depend on human judgement can bring the greatest benefit. Reviews by operational risk and internal audit groups can make sure process change stays on track. And whilst RPA has a technology component, it should have a strong business sponsor and not be led by IT.

Rather than just freeing up a few hours in each staff’s week, banks will obtain a far better return on investment if they entirely free up the work of FTEs they budgeted for and change activities across roles so that work can be automated more consistently. In order to scale RPA, banks also need to look at it in the context of their overall operational strategy in order to make sure they don’t get stuck with antiquated processes or systems that they intended to phase out.

And despite the cost reductions, Liew said “we have not large scale retrenchments.” Instead, banks are “redefining the job scope, freezing headcount, so people plus robotics handle higher volumes without more people.” Eventually, though, there will be pressure to look at letting go people.”

Case studies
ICICI Bank
“We started the journey about 18 months ago,” ICICI’s Anita Pai told The Asian Banker. “When we tracked the flow, we found our teams are busy inputting data, which left little time for evaluating data quality or other analytics. When we were looking at budgets and saw the growth, we looked at how can we get a breakthrough. That’s when we figured out there are software robotics.”

The focus is on high-volume rules-based processes in retail operations and now includes 200 processes, Pai said, which cover 10% of transactions in the back office. “We put up plans to double it by the end of March this year. We are on track,” she added.

As an example, the bank receives complaints from some customers who go to the ATM and claim that they didn’t receive any cash, even though their account was debited. Whereas it used to take three to four hours or more to resolve the complaint, customers now receive a response in less than one hour.

A robot checks at the customer account level and at the switch level, reviews the ATM e-journal, checks whether the customer is a habitual complainer and then makes a decision about whether there was an error. Customer onboarding is another example, with sub-processes undergoing automation including data entry, verification, address matching and checking identification.

The bank is now completely volume-agnostic in processing requests said Pai and accuracy has increased because robots don’t make mistakes such as mis-reading a “7” as a “1”.

The benefits of using RPA including processing a far higher volume of transactions more efficiently, increased accuracy, and customer delight. While ICICI Bank had been adding several thousand employees a year, it may not need to add as many new staff as before. Moreover, Pai said that the impact on staff has been positive because repetitive work has been taken away. “Staff are happy. They leave office on time,” she quipped. One advantage ICICI Bank has is that the average age of staff is young, at 28 to 29 years of age, so there has not been an issue of not being able to upskill staff. “We have internal job postings. These people move into different areas. People are happy to explore,” she explained.

Along with the advantages for staff, customers have benefited as well. “Customers have started seeing an improvement,” Pai said. Instead of feeling anxious when they call the call centre, customers now get an answer without even needing to call. “There has been an impact on the customer,” she said, even though the customer may not know that it’s due to robotics.

Danske Bank
Danske Bank has developed a robotic centre of excellence (CoE) which it said can deliver robotic solutions through a streamlined factory setup within six to eight weeks. The bank said that it is using RPA to “automate current tasks as if a real person were doing them across applications and systems,” making sure that Individual delivery teams or process excellence resources can do it without extensive IT training or help. The results include a 45% increase in increase in ability of employees to focus on customers, a 40% reduction of average process execution time and payback within three months after deployment.

RPA was used in the unit which handles remittances from third parties for customers who have paid part or all of their debt, for instance, which handles everything from bookkeeping and updating the limitation date to editing the risk profile and handling follow-up. Pressure on staff has been reduced, the time for each case has been decreased from 90 seconds to 38 seconds, expenses were lowered and the quality of the output has been high.

The keys to success, according to Danske Bank, include starting small, being realistic, having robust monitoring of the robots, being ready when robots fail, and communicating success.

An effective solution to current challenges
At a time when banks face increasing pressure to deliver ever-higher service levels and innovative solutions while also reducing cost and increasing efficiency, RPA offers an effective solution. While the greatest benefit may be in higher-cost developed markets, banks in lower-cost emerging markets can benefit tremendously as well. And it’s important to note that the benefits go beyond just cost, as customers get a better experience while staff can focus on higher-value work. Implementing RPA sooner rather than later, and across as many processes as possible, is essential for competing effectively in an increasingly dynamic retail banking environment.