By W.B. King
During a recent episode of Multimodal’s podcast Pioneers, Wescom Financial Services President Jonathan Allen and Wescom Credit Union’s CIO Joseph Pellissery shared how they have successfully combined robotic process automation (RPA) and generative artificial intelligence (AI) solutions to streamline back-office processes, while elevating member interactions.
“Before LLMs [large language models], ChatGPT and Generative AI, we invested a lot in RPA, which was the known form of AI 10 to 15 years ago. Over recent years, we have been able to save over 25,000 manual hours and counting using RPA,” Pellissery told podcast host Ankur Patel, founder and CEO of Multimodal, a New York City-based generative AI company focused on automating complex workflows in insurance, banking and healthcare.
“As technologists, it’s our responsibility to bring the possibilities to the table—how a particular technology can be translated into a business value,” Pellissery added, noting that Wescom Financial Services, a CUSO, is the wealth management arm of Wescom Credit Union.
“If you look at the broader ecosystem, every company claims to be an AI company, but how do you generate value?” Pellissery posited. To determine an answer to this question, the organization recently partnered with Microsoft and its 365 Copilot program, what Pellissery called an “economical approach,” from a cost perspective that solves “for some of the low-hanging fruit use cases” for generative AI.
According to Microsoft’s Chairman and CEO Natya Nadella, the solution combines the power of LLMs with an organization’s data in the Microsoft Graph.
To test the waters, the $6 billion Pasadena, Calif.-based Wescom Credit Union, which supports more than 200,000 members and 30 branches, implemented a 365 Copilot and Microsoft Dynamics proof-of-concept contact center solution.
Over the course of a year, for example, Pellissery explained that the contact center can field many thousands of calls. By selecting a subset of 1,000 calls and studying corresponding transcripts, generative AI was able to identify only those calls deemed problematic.
“The quality assurance agent can ask questions: What was the emotion of the call or was the member frustrated? So, the quality assurance process is now disrupted in many good ways,” Pellissery noted, adding that he also oversees payments and digital solutions at the credit union. “They have a lot more time now to focus on coaching and conversations that would improve the team member experience and improve the member experience.”
Bridge the Tech Talent Gap
As a division of Wescom Credit Union, Wescom Financial Services offers full-service investment options through its broker dealer CUSO Financial Services (CFS), which helps Wescom members meet financial and investment goals, explained Allen.
“My role with the organization is not only as president of Wescom Financial Services, but I also have responsibility for overseeing our property management casualty business,” he said. “We have a personal lines business and a very compelling insurance agency and additionally, I oversee the deposit strategy for the credit union as well.” Allen added, “We need technology-savvy advisors to bridge the talent gap in financial planning for the next generation.”
Also under its business umbrella is Wescom Resources Group. This tech-focused CUSO serves 105 other credit union clients with a combined $148 billion in assets, which serve more than six million members, explained Pellissery.
“I think it is an incredible vantage point because you get to see the problems across the space—what are the common threats, what are common needs and how do you serve it,” Ankur noted. “You’re basically solving for enterprise set of problems but across a set of credit unions that operate independently from one another, but have that shared need.”
Pellissery explained that while the CUSO is working with Microsoft, it is also currently considering a partnership with Amazon Web Services. “For our RPA solution, we have also looked at UiPath and Kofax,” he continued. “It’s a lot of partnerships with industry experts and vendors and then we add on the value from our team based on what is specifically an extension of the standard capability.”
The Foundation for Integration
When Ankur asked how RPA successes compare to next generation AI technologies, Pellissery said that RPA best serves repetitive and predictable business manual use cases.
“When there is a predictive nature to inferring the next step—deviate from step two to step six, where you have to predict what your next step could be based on patterns, based on LLMs—that is where generative AI and machine learning capabilities are best used,” he said.
Before considering the capabilities of RPA, generative AI or machine learning solutions, Pellissery said an organization must have a robust data and application program interface (API) strategy.
“That’s the foundation—what you build on with RPA, generative AI and machine learning. Having your historical data in place becomes critical,” he said. “Make sure the foundation for data, the foundation for integration is set in place before you embark on some of these newer capabilities that are being delivered today.”