From the February issue of UCD
For lenders like Kevin Lawson at First Texas Auto Credit in Coppell, Texas, every minute in making a credit decision matters. That’s why he and many other lenders are moving forward with ambitious artificial intelligence plans in the new year to help with underwriting and collections.
“We anticipate creating more efficiency in the business and see how fast we can fund,” Lawson said.
Several leaders of lending institutions and financial companies in the F1 20 Group disclosed their efforts to incorporate more AI into their workflow at their January meeting in Irving, Texas.
As Lawson expressed, AI is anticipated to help with efficiency, especially with underwriting. For contracts that will definitely be funded, AI can sort through paperwork to make sure all the appropriate forms are provided and make a quick decision, the same as larger companies.
“If a contract comes in from a dealer, AI can take the 30 pages in the PDF and separate them into different buckets. If it hits all the set parameters, it’s ready to fund, and we could tell the dealer with AI,” Lawson said. “Up to 20 percent of our contracts would fund quickly. Even if they are missing a form, we could take an hour to get a contract.”
Several lenders are working to train large language models or are working with vendors to train AI tools to look for certain items.
“We’re working to drill down to everything wanted for an auto-decision tool for underwriting,” said Chad Brown of Tracir Financial in Reynoldsburg, Ohio.
Ben Stefanovski of Tracir Financial added that AI training includes looking at the positive and negative attributes that are part of the scoring for decision-making. AI tools will use an algorithm to assess risk and then provide a score.
The efficiency of the AI agents and tools is also expected to help the lenders expand their business. With AI handling the contracts of the deals that will definitely be approved, the lenders can have their staff give extra scrutiny to those deals on the fence. The staff can ask for extra documentation and clarification to make more informed decisions as they are freed up.
“You always need the human element to make decisions. This will all you to reallocate resources,” Lawson said.
The lenders stressed that human oversight will never be fully replaced.
“It’s a long-term investment for auto-decision on the highest scoring loans,” Stefanovski said. “But there will always be a hybrid for those mid and low scores where you need the human element to decide.”
For accounts already funded, lenders are looking to AI to enhance their collections efforts, handling phone calls and inputting data.
Collin Grabowski of Gateway Financial in Saginaw, Michigan, said he is understaffed by 18 collectors and is looking to address the shortage with AI. Grabowski said the move is to help with efficiency and expenses.
“A lot of the day is spent inputting data. We want to move collectors to make more of an impact,” Grabowski said.
Several lenders are looking to AI to handle the lion’s share of the calls, pointing out how voices can be adapted for different regions.
“Around 50 percent can be handled by AI. It’s the really odd calls that AI can’t handle,” Charles Gaver of Heritage Acceptance Corporation in Elkhart, Indiana said.
Another advantage is the ability to take calls after hours and weekends to have data ready to be followed up on the next business day.
Work continues to implement the AI into their operations, but the lenders are hopeful to see results by the end of the year with servicing more loans and growing their businesses.
“I’m the most excited that I’ve been in my entire career,” Lawson said. “I’m not a tech person.”
