Automation Must Be End-to-End

Speakers at the DTCC Fintech Symposium held in New York point out that automation is crucial for the industry as it deals with an influx of data. Around 90 percent of the data that now exists was created in the past 10 years, according to Roger Park, partner/principal and strategy and innovation leader for financial services at the consultancy EY.

“Today, there’s a pervasiveness of data, just a lot more data that we’ve ever seen before,” Park says.

As a means of combing through all of the information now available, automation technology—whether it’s robotic process automation, machine learning or artificial intelligence—is also now more pervasive. Park says costs of these technologies have gone down so financial institutions have more access to them.

But automation projects are not holistic, as some institutions prefer to keep automation to only one process—onboarding, for example—but it is not deployed elsewhere. And that, speakers at the forum say, is where the problem lies. DTCC managing director and CTO for enterprise infrastructure James Lee says some automated systems are used only by a few processes, which can limit the amount of data that is processed.

“The way we’ve approached this before is in a silo, and that’s where the challenge is,” Lee says. “You automated specific tasks as opposed to a holistic end-to-end business aspect, but because of the data explosion we can’t do that.”

He adds that firms that want to go the automation route should look at it from a value standpoint and ask how the technology can be better used.

Automation simplifies not just how information is handled, but how transactions are analyzed and recorded. Keeping only one area of the business automated makes little sense. For Lee, the big issue facing financial firms when it comes to automation is the lack of interoperability between systems and processes.

Unfortunately, some financial institutions may still find themselves unable to fully automate all processes, or prefer certain tasks to remain manual. But if any process is automated, it makes sense to automate all.  

Christopher Surdak, program director at the Institute for Robotic Process Automation, says thinking about new ways to look at investment costs could be the answer to fully automating business processes, even those that seem resistant to change, which he calls “superbugs” of the IT world.

“We tried to automate that step for 45 years and it hasn’t been done yet. The reason could be cost, complexity, or regulation,” says Surdak. “If you measure that technology with this superbug and still use the same return on invest model you’re going to have to think again. To get this right, you need new metrics.”

He says firms, again and again, are unsure of how to make pilot projects fit into their whole ecosystem, mainly because they insist on going back to return on investment models built to track systems like Excel sheets rather than disruptive new technology.

EY’s Park notes that despite the challenges—and some stubbornness—of firms, automation is now on the cusp of becoming even more widespread.

Efficiency offered by automation also means firms, and their employees, have to be more unique. Bank employees must also prove that they can think out of the box. Automation can do a lot of the jobs people do except for idea generation or working without certain parameters set.

Surdak says banks and their employees need to differentiate from answers or tasks a robot can do. He says it is the “crazy” employee, the one who comes up with out-of-the-box ideas and solutions, that is complementary to the automation process.

It can also change the way some banks outsource certain services. Business process outsourcing is directly threatened by automation since it eliminates the need for requiring a larger team to handle specific tasks. 

Surdak, Lee, and Park had one last piece of advice for financial firms looking to automate: Start smart and focus on the business value of the project and converge it into different platforms to increase its interoperability.

BOTTOM LINE:

• Automation shouldn’t be done in a silo but rather in a holistic way throughout the organization.
• New metrics can encourage automation for “automation resistant” steps.