As a new feature of our site, we’ll begin highlighting Wescom’s innovation based on interviews with its leadership team. First up is an interview with Desigan Reddi, Vice President of IT Operations and Architecture at Wescom Financial, who leads Wescom’s efforts on AI empowerment.
These days, it’s hard to identify an industry or aspect of human endeavor that is not being impacted by artificial intelligence. Financial services is a case in point.
A 2024 study by Cornerstone Advisors revealed that nearly 30% of credit unions plan to invest in AI chatbots this year, while 42% of community banks are still in the discussion phase, and 24% have not yet considered the technology.
Within just the past few years, artificial intelligence has risen to a position of ubiquity in our daily lives. Large language models (LLMs)—a subset of generative AI technology trained on massive volumes of text-based data to understand and generate output in human-like language—have captured the imagination of the public. Today’s consumers use LLMs like OpenAI’s ChatGPT, xAI’s Grok, and Google’s Gemini for everything from writing resumes to making travel plans.
Yet, the pervasiveness of LLMs has clouded consumer expectations in a way that’s analogous to how the public became comfortable using Google search. Consumers gradually came to expect that same level of instantaneous, always-on responsiveness from all service providers, including their financial institutions.
“I think that’s the first hurdle we have to cross,” says Desigan Reddi, Vice President of IT Operations and Architecture at Wescom Financial. “The industry is inundated with providers that are AI capable. There has been a flood of technology, but from a practicality perspective, few vendors have been able to leverage it for real world use cases.”
That’s because practical use cases in financial services are more discrete, complex, and challenging, as they come with onerous regulatory compliance burdens, data security requirements, and multi-faceted, integrated workflows.
For these reasons, simply repurposing or customizing popular LLMs like ChatGPT to meet a credit union’s complex needs is exceedingly challenging.
“LLMs give you generic general information based on years of consuming publicly available data on the Internet, versus the proprietary data that is only pertinent to your business,” Reddi adds. “Programming an AI engine to work with that proprietary data is not as turnkey as everybody thinks it is.”
For Credit Unions, Back-office Opportunities Abound
This is not to say that financial institutions haven’t been actively testing AI in a variety of real-world scenarios.
According to research from Jack Henry, 54% of bank CEOs and 41% of credit union CEOs named efficiency as a top strategic priority – marking the first time in the firm’s annual study where efficiency has taken the top spot across total respondents. Advanced cognitive technologies—including generative AI, machine learning, optical character reading (OCR), robotic process automation (RPA), and workload automation—have the unique ability to streamline and enhance operational processes, giving time back to organizations to repurpose toward higher-level, more strategic goals and efforts.
Credit unions and banks have used chatbots to provide immediate responses to member inquiries for years. However, according to Reddi, most of these chatbots are powered by old-fashioned rule-based workflows, rather than true generative AI technology.
“A lot of members think they’re dealing with an AI bot, but most of the bots people interact with, especially in our industry, are still the old-style workflow,” Reddi explains. “They use either old-style calculations or an AI engine to understand intent. But once intent is understood, you’re then routed to a traditional workflow bot with prescribed answers and not real generative AI where the answer is generated by the AI engine itself.”
Greater advancements are being seen in the back-office, where AI-enabled solutions have proven useful in a variety of use cases, including loan processing and underwriting, document management, fraud and risk mitigation, personalization and predictive insights, and workload automation.
Enhancing Operational Excellence With AI
AI is also proving adept at enhancing operational efficiency in member-facing departments like the contact center. For example, Wescom Financial has deployed agentic AI technology—artificial intelligence systems that can make decisions and act with minimal human interaction— to automate member call summaries.
The project began soon after the public release of ChatGPT, when Reddi’s team was seeking innovative use cases for the technology.
“Traditionally, a Member Contact Center (MCC) Representative was responsible for manually summarizing each member call,” Reddi says. “And we noticed that representatives would write in shorthand and three letter acronyms. It became really difficult for members of other departments to decipher exactly what happened on the call, unless you knew the lingo.”
Reddi’s team began testing ChatGPT to summarize the call transcript down to 100 words or less and disposition the call. The process took about five seconds and was reasonably accurate.
“That has now evolved to where we’re doing this in real time,” Reddi says.
Once the call is complete, the MCC Representative presses an agentic AI button, triggering an immediate summary of the call the representative can view. The AI agent has been trained on specific categories and terms that both the MCC and other departments understand. In just five seconds, the AI agent provides a top-level disposition, a sub-level disposition, and a 100-word summary. When compared with the time it previously took for a human agent to manually summarize the call, Wescom has recouped roughly 25 seconds of non-call time from each interaction.
Wescom is also using agentic AI to streamline and enhance quality assurance (QA) screening in the contact center.
Previously, Wescom staff would review only five percent of the recorded calls that came through the MCC. This proved to be an arduous and time-consuming process, as the reviewer would listen to the entire call, then manually answer a 15- to-20-question survey to determine whether the call was handled properly.
Now, with the help of agentic AI, Wescom can review every call to the MCC (roughly 40,000 per month) and rate each one on a five-point scale. Any calls that rate poorly are sent on to a human QA agent for further review and follow-up.
As a result, Wescom has noted improvements in member satisfaction through both MSAT and NPS (Net Promoter Scores).
Take a Measured Approach To Deploying AI
For credit unions considering testing the AI waters, Reddi recommends taking a measured approach.
“Given that our industry is so member-focused, we don’t want to risk deteriorating that member experience at all,” Redding says. “We’re very hesitant to put something in front of our members until we’re absolutely confident.”
With this goal in mind, Reddi’s team tests all new AI-enabled capabilities with credit union staff before introducing them to members.
Currently, Wescom is developing new use cases like an AI agent that assists in member case management. The agent helps craft a response email and identifies potential solutions based on previous case history and a proprietary knowledge base.
“All of that is happening with moderation and with human eyes on it,” Reddi says. “Once we see a good efficacy on the responses we’re getting from generative AI, then we’ll consider deploying it to our members.”
Charting a Course to an AI-enabled Future
A mantra Reddi follows is: “AI is not going to replace the job, but the person who knows how to use AI will replace the job.”
Just as email replaced interoffice memos and faxes decades ago, today’s workers will need to understand and embrace AI-enabled workflows as the inevitable future.
“We all need to figure out how we can leverage technology to upscale ourselves and remain relevant,” Reddi says. “There’s apprehension about AI replacing a lot of functions in the workplace, but the way we’re approaching it is that it’s going to create time for our team members to focus on the important things — member service and the human touch — which the credit union industry has always prided itself on.
“If we can create an extra hour each day to allow someone to make a meaningful connection with another human being, I think that’s important.”