Not your father’s robotics: what you may not know about bots in financial services

By Gordon Sands, Executive Director, IT Architecture, DTCC

The word robotics often conjures an image of a machine on an assembly line, tightening screws or sealing boxes over and over endlessly, but today’s reality is far different. Artificial intelligence-driven robotics can supercharge massive databases by accelerating the capture, analysis and distribution of their data and enabling them to continually learn and improve. For back-office applications, robotic process automation (RPA) can replace repetitive data-entry tasks with automated processes that can accelerate workflows and slash error rates.

Not only has the automation of repetitive processes moved from the factory floor to the computer desktops of workers in a range of business sectors, it has expanded organisations’ bandwidth, enabling them to focus on higher-value activities that improve outcomes for employees as well as clients. And financial services firms are leading the way in incorporating robotics in our operations.

Top 3 myths about robotics

As companies in the financial sector, including DTCC, have implemented robotics projects and put new robotics pilots in motion, our experiences are helping dispel some lingering misconceptions about the use of robotics in business. Top among them:

• Robotics is a jobs-killer.

• Robotics can only be used tactically in business, to automate simple, repetitive tasks.

• Robotics introduces unmanageable risks to an organization’s data and security.

As financial services firms have learned, these notions are generally misleading and outmoded. Real-world experiences can demonstrate how robotics is deployed and affects organisations’ human capital and applications development processes.

To note, RPA refers to computer software that is configured to automate repetitive, deterministic processes, whereas “smart” robotics can deliver even greater value by augmenting RPA with artificial intelligence (AI) and machine learning.

Fact: Robotics can increase employee and organisational bandwidth

While robotics, just like other advanced technologies, certainly can be used to reduce fixed costs, job-cutting is generally not the rationale behind the push into robotics and RPA in the industry today. Rather, firms harness RPA to expand their capacity, by freeing up staff from rote and repetitive manual tasks, and to augment and improve the quality of the work people continue to do, by reducing manual error in low-end data entry and workflows, extending the coverage of operations to address a wider range of transactions, or speeding up the work itself.

RPA initiatives are particularly valuable in automating processes that entail large volumes of data and have traditionally required some amount of manual effort. In these cases, the size and availability of staff assigned to handle the work can limit how quickly and thoroughly the data can be processed.

By removing these constraints, RPA gives firms additional capacity in return for minimal up-front programming costs. And with the right tools such as specialised RPA software, these initiatives can typically be implemented in short timeframes.

Some firms are using RPA to simplify their Know Your Customer (KYC) and client onboarding systems. Robotics can help speed the multi-stage data verification and hand-off processes financial firms require for new-client intake.

Finance departments, with their frequent accounting, billing and reporting processes, have also been proving grounds for a number of RPA projects. Automation has been able to boost the accuracy and timeliness of procedures like billing reconciliation. Traditionally, due to capacity limitations, reconciliation processes that seek to ensure figures are consistent across two or more systems sample only a portion of the total to confirm accuracy. But with RPA, the entire population of figures can be checked at rapid speed and any non-matches flagged for investigation. The resulting robot-generated reconciliation files are instantly auditable, eliminating the need to prepare separate audit reports. And staffing and organisational capacity is expanded when other finance functions are streamlined and enhanced.

Experience has shown that employee satisfaction rises when RPA is implemented, because employees are freed from rote tasks to do what they do best: focus on higher-value, more-creative analysis and troubleshooting of those cases that fail to reconcile. And the entire organisation benefits from gaining bandwidth.

Fact: Robotics combined with AI can be used strategically

While RPA initiatives are proving to be excellent tactical measures for automating repetitive, data-rich processes, increasingly firms are employing advanced robotics for strategic purposes, to increase market share or build out new product or service offerings, for example. In these cases, robotics is often augmented by companion technologies like AI and machine learning to expand the bot’s capabilities and create material competitive advantage.

Enormous databases that change dynamically and require frequent updating can be ripe for enhancement through robotics, AI and machine learning. Many firms are seizing the opportunity to leverage recent advances in cognitive technology and data mining to expand their databases’ data capture and self-learning capabilities, to automate data sourcing, boost the number of data points, and streamline clients’ data sharing and collection.

The strategic benefits are significant: databases will capture higher-quality information and do it more frequently, push it out to the industry faster and continually improve by incorporating user feedback. And, by alleviating the standard burdens of manual data processing, firms acquire new capacity to perform data analysis, a boon for the companies and their clients.

Fact: Robotics applications, when properly programmed, can mitigate risk

Maybe people have watched too many science fiction movies, because robotics is commonly misunderstood to put our organisations at nearly boundless risk. Scenarios portray out-of-control bots running amok, exposing private, confidential data; making a process go into overdrive or, conversely, shutting down completely; erroneously multiplying transactions a million-fold or inexplicably wiping them out. 

Incidents like these are extremely unlikely because bots don’t have minds of their own and cannot act autonomously. Bots are pieces of software, albeit faster and higher-capacity than earlier generations of code, that execute only the actions humans program them to do. Even smart bots – robotic processes augmented with AI and machine learning – are not human; they are still subject to the constraints we impose on them. We can and should always program them narrowly to restrict their functionality, limit their operational run-time, and tightly circumscribe their entitlements to the data and to other systems. If we design, code and test our bots properly, they are no more likely to misbehave than any other software applications. 

In fact, bots can mitigate the risks of processes going haywire. Unlike with humans performing the same tasks, we can impose even stricter controls and permissions on the activities of robotic users, thereby improving an organisation’s risk profile. Bots won’t mind being monitored and checked continuously, and they won’t lose their concentration, fat-finger a keyboard, or misconstrue their instructions.

Nevertheless, firms shouldn’t minimise the challenges of integrating robotics into their existing applications and infrastructure because interactions between bots and other software can create unexpected problems. For instance, if a firm tweaks an application like email without considering the knock-on effects, bots touching that application can be disrupted and prevented from executing successfully.

This lesson is often learned when a familiar software tool crashes without warning; people are good at making adjustments on the fly, but without careful planning and error handling, bots just stop working. But such experiences can help firms refine their operating models to incorporate improved change-management practices and promote open-source sharing of the bot interfaces with underlying applications such as email, Excel and Salesforce.

Transforming data management

Advanced technologies like robotics and AI are not the only tools for boosting efficiencies and reducing risk for financial services organisations, but they will be the leading drivers to transform firms’ data and process management in the coming years. That’s why more and more companies will integrate these technologies into their operations, through RPA initiatives and the application of AI and machine learning.

These efforts will present operational challenges as more and more processes are automated and interlinked. But the results will enable financial services organisations to generate “smarter” deliverables and continually improve service to clients.