Automation and AI are increasingly deployed in the financial sector, cutting time for repetitive processes

Automation in the finance industry is cutting time spent on routine tasks, and becoming increasingly important to operations—but the security questions around the deployment of the technology are always a concern.

Osaic, a wealth management fund previously known as Advisor Group, started its automation process around 15 months ago, according to the company’s VP of digital automation, Matt Ham.

Osaic’s customer base had been unsatisfied with the financial services company’s level of technology, Ham told reporters at Automation Anywhere’s Imagine conference in September, which motivated the firm to incorporate AI in its processes.

Numbers game. The company quickly found that using software robotics added value to operations and streamlined processes. By early this year, some 42 status bots delivered 17 to 20 labor hours of value a week. The increase in efficiency began paying for itself—and at a financial services firm, profit’s the goal.

“We’re looking at our solutions from an end-to-end business perspective, not just looking at it from an IT perspective and trying to fit the business into the IT,” Ham said.

He added that Osaic has expanded software robotics to help validate information for ACH transfers; the bots are “essentially doing the work.” There’s still work to be done—he gave as an example making sure decimals are in the right place for payments and deposits—but keeping a human element for oversight and control still results in a more efficient process.

At KeyBank, service digitization domain manager Mike Reynolds helped start the automation process back in 2017. The company trained 45 people across essential operations and moved 22 automations into production, he told IT Brew.

Robotic process automation, commonly known as RPA, does about 70% of the work for Reynolds and his team. Using software robotics can automate up to 20% of their banking controls.

Take care. Implementing automation does come with some risk, Reynolds acknowledged. Common sense guardrails help manage the danger.

“From a security standpoint, there are always applications they’re monitoring, desktops,” Reynolds said. “So we have to really get those preapproved and say, here’s a new install.”

Automation Anywhere CIO Sumit Jahor told IT Brew that in order to properly deploy AI, vendors and customers need to be on the same page. That includes security, where clarity is essential.

“Data security, data privacy, selecting the right model, training, tuning, prompting, the trust layer…is actually going to make it easy for an automation person to be able to deploy it without knowing everything about gen AI and how it functions,” Jahor said.