Auto lenders have been approaching generative AI with caution, spurred by regulatory oversight and limited use cases, but there are lesser-known applications financiers can implement as the technology improves.
While the functionality of generative AI is advancing, the tech is not fully ready for lenders to implement into back-end operations due to flaws, Tom Oscherwitz, vice president of the legal department at fintech Informed.IQ, told Auto Finance News.
However, lenders can use generative AI in its current state to build code, generate ideas and provide personal assistant capabilities, Oscherwitz said.
“When you’re building code, you’re likely not the first person who has ever done that code,” he said. “If you have a giant library and people have done similar projects, you can get that code.”
But lenders must use generative AI appropriately, cautions Justin Liebstueckel, senior manager at Porsche Consulting.
Gen AI as ops tool
“Lenders have to use generative AI as a tool in their toolbox of many, and you have to do your homework with regards to technology to really enable generative AI to do what it’s supposed to do best,” Liebstueckel told AFN.
The greatest use case for lenders now is utilizing AI to train company software for use in customer service centers by training the AI using the company’s internal data, he said.
“Think about [generative AI] as a tool that you can train on all the knowledge that you have in your company as a lender, all the regulations in place and all the knowledge that is passed from generation to generation of agents,” Liebstueckel said. “Considering the turnover in these kinds of service centers — 15% to 20% — generative AI will help address the knowledge transfer and the capabilities of an agent communicating with a customer.”
Once trained, generative AI could be used to help a customer amend a contract or propose a change to the lender’s internal team. A human employee could then authorize the change and send it to the consumer, Liebstueckel said.
AI “can also translate into other parts of the service center such as funding, credit approval funding and servicing of contract collections,” he added.
Chase Auto, for example, uses AI for coding, to shorten funding times through automation and to monitor fraud.
Lenders can also use generative chatbots in customer communications to answer consumer questions without human interaction, Informed’s Oscherwitz said. If there is a question the chatbot can’t answer, the consumer is directed to a human representative.
Current use cases are narrow, but they can still elevate employee work before the software reaches its full potential, Oscherwitz said.
“With generative AI, we’re getting closer to the concept where we tell it, ‘Make this form, design this program, develop this compliance policy,’” he said.
Gen AI can impact residuals
Generative AI can impact vehicle residual values, and lenders should take a close look at how manufacturers are using it in their vehicles and the impact on residuals, Liebstueckel said.
For example, Mercedes-Benz implemented ChatGPT into its in-car voice assistant capabilities, he said.
The acceleration of in-vehicle technology and software updates “confirms that lenders will need to know more about the cars, the technology that is in vehicles and the software that is running on the cars when it comes to resale value,” Liebstueckel said.
Without that knowledge, a lender could be missing out on profit “by not knowing exactly what the car’s worth, by not knowing what software is in the car, what kind of hardware is in the car and what is the car capable of that might not be part of the options list,” he said.
