By W.B. King
Persistently focused on member-driven growth, 4Front Credit Union, which supports more than 100,000 members and 18 branches, recently recognized that its credit decision making process needed an overhaul to keep pace with competition.
“Smart and simple banking is our aspiration. Like everyone in the financial space, we’re continuously challenged to improve speed without sacrificing quality and improve ease without sacrificing risk measures,” 4Front Credit Union’s Chief Lending Officer Daniel Baker told Finopotamus.
The $1 billion Traverse City, Mich.-based credit union, which nabbed the 2023 Tekkie for Lending, also won the 2022 Tekkie Award for Marketing and Member Communication.
“One area we honed in on was the speed at which we deliver our loan decisions and the quality and consistency of how we come to those decisions. Finding a solution to help increase our total decision automation and, more importantly, our automated approval percentage was a high priority as we targeted a faster, more data-inclusive decision process,” Baker continued. “We also identified the need to remain consistent in our decisions, which was emphasized by recent geographical growth for the credit union. We knew we needed to find a strategic partner to help us be fast, efficient and consistent.”
After vetting three other vendors in late 2021, the credit union selected Scienaptic AI as it “checked all the boxes,” Baker said. The New York City-based fintech offers an artificial intelligence (AI)-powered credit underwriting decisioning solution.
Prior to the partnership, Scienaptic AI’s Chief Marketing Officer Chandan Pal explained that 4Front Credit Union relied “too heavily on bureau data and simple rules,” which, in turn, limited its decisioning capabilities.
“To succeed, 4Front Credit Union needed decisioning models adapted to its unique growth needs and risk appetite,” he told Finopotamus. “The credit union searched for a decisioning platform that could integrate seamlessly with its previous setup and provide scalability as its needs evolve.”
The leading challenges, he added, were:
Bringing automation to its existing manual underwriting process.
Utilizing multiple sources of data, including core data, loan origination system (LOS) data, traditional bureau scores, and alternate data sources.
Deploying decisioning models adapted to their unique growth needs and risk appetite.
Seamless integration with their MeridianLink Consumer LOS.
Team Work Works
In tandem with its new loan origination system integration (MeridianLink Consumer LOS), from contract signing to rollout, Scienaptic AI’s solution took between 90 and 120 days to go live.
Approximately 40 of those days, Baker said, “were reserved for live testing as we let our decision strategy run behind the scenes and provided us with data to validate the initial decision parameters and the quality and performance of those decisions.”
While he noted that it’s easy to become “paralyzed in analysis,” and conceded that he and his team like to “nerd out” on data, there were important lessons learned.
“Underwriting philosophy is a point of pride for 4Front, so there came a time when we needed to accept that the solution design was capturing how we view credit and can learn as we evolve,” he continued. “Trust your data validation, and don’t let pride stand in the way of efficiency. These systems are highly configurable and will continue to deliver actionable data for future adjustments to strategy.”
Implementing the solution, he said, was "simple but not simplistic," noting that there were many moving parts, including the noted LOS integration. Baker added that the credit union’s Vice President of Lending Nate Peterson and lending program administrator Nathaniel Rodriguez worked diligently to ensure 4Front Credit Union’s philosophy was captured.
“Having Scienaptic’s guidance was invaluable, paired with their openness and desire for constant feedback as we navigated the project,” he said. The team at Scienaptic and our team at 4Front worked together exceptionally well, making a complex solution rollout look and feel easy.”
Game Changing Results
Scienaptic AI’s lending platform, Pal explained, includes “cutting-edge” AI and its “intelligent decisioning cloud” as well as pre-configured data application program interfaces (APIs). Additionally, the company offers a LOS integration kit for seamless integration with existing origination systems.
“By incorporating data from multiple sources, Scienaptic’s self-learning AI algorithms allowed 4Front Credit Union to identify credit-worthy applicants with lower bureau scores and thin files, making lending more accessible to underserved members,” he said.
The results, Pal added, have been “game-changing,” including:
Automated accept/reject/review decisions allowed 4Front Credit Union to process loans faster than ever before.
Explainable AI models increased auto-approvals in a fair and transparent manner.
Seamless integration with MeridianLink Consumer created a unified end-to-end loan origination, processing, and decisioning setup, resulting in an unparalleled level of efficiency and speed.
“Since launching the solution, our loan decision automation has increased from 18% to 56%,” Baker told Finopotamus. “We are experiencing a consumer loan net promoter score of 86 year-to-date and have increased our approval, funding speed, and loan products per member.”
When asked if he had advice for other credit unions looking for a new lending solution, Baker said if they are currently in the discovery phase, the credit union should realize it has underutilized data and could perform better credit underwriting.
“I advise ensuring you find a strategic partner. Be unapologetic about asking tough questions, voicing every aspiration your organization has for the solution, and pushing to corroborate their answers,” he offered. “The result will be a partnership and a delivered solution that exceeds your credit union’s and your members’ expectations.”