AI agents address KYC, retrievals, underwriting in equipment finance

Practical applications of AI agents are taking shape in the equipment finance and insurance industries while tech providers push the bounds of their capabilities.  

The agentic AI market is projected to grow nearly 1,622% to $107.3 billion in 2032 from $6.2 billion in 2024, according to research and consulting firm SNS Insider. AI agents — designed to carry out tasks autonomously by perceiving their environment and taking initiative with minimal human oversight — have begun to transform equipment finance by speeding processes and lowering costs.  

Equipment lenders should fully “dive in” to unlock the potential of agentic AI and enable entirely new business models, Bill Verhelle, chief executive of embedded lending platform and equipment finance service provider QuickFi, told Equipment Finance News.  

“In the long term, adding AI agents to the existing, 60-plus-year-old business model won’t get legacy companies out of the woods. They must contemplate how the business model will change with these revolutionary, emerging capabilities. And they will need to find a way to get there quickly. That is a much bigger challenge facing all banks today.” 

— Bill Verhelle, CEO, QuickFi

AI agents for compliance 

Nearly 80% of more than 500 surveyed financial services firms were victims of payments fraud in 2024, according to the Association for Financial Professionals. For equipment lenders, adhering to Know Your Customer (KYC) and Know Your Business (KYB) compliance requirements is crucial amid increased fraud and digital adoption.  

Using agentic AI is an effective way for lenders to meet these requirements, Rohan Marfatia, chief executive of data and technology provider DataCRaiM, told EFN.  

By automating end-to-end processes, a lender or legal team member can request agentic AI to pull data from a Salesforce package and confirm a business’ name and taxpayer identification number (TIN) to verify identity, he said. 

“And then the agentic AI — it could be Agentforce, it could be Copilot in the Microsoft ecosystem — it’s able to retrieve the data, state that the TIN was verified,” he said. “Whereas previously, a human being used to do that function. You had to manually go into that managed package or into that function, click on certain buttons and then extract the data. So, that manual task got replaced by an AI agent.  

Agentforce is a Salesforce platform designed to deploy AI agents, powered by the “Atlas Reasoning Engine,” which uses reinforcement learning, feedback loops and other advanced techniques to understand complex service queries, generate accurate responses and drive autonomous decision-making, according to Salesforce.  

‘Messy inbox problem’ 

When mulling potential applications of agentic AI, equipment lenders should consider time-consuming manual processes that require judgement and analysis, also known as the “messy inbox problem,” QuickFi’s Verhelle said.  

In equipment finance, these processes include loan applications, underwriting and documentation, he said. Agentic AI is also well suited for “automated workflows within embedded lending.” 

“The entire embedded lending customer value chain, from application to final payment, can be improved and highly automated with agentic AI. This is not the case with manual processes, where a salesperson, a credit analyst, a documentation clerk, a funding manager, a collector or a customer service agent undertakes manual actions and types notes into a CRM.” 

— Bill Verhelle, CEO, QuickFi

In addition, AI agents could be a boon for equipment insurance providers, Verhelle said. QuickFi has created an AI insurance agent that has capabilities including opening and reading emails, filing certificates, documenting certificates and identifying deficiencies in the certificate, he said. 

“If it’s not the right coverage amount, not the right machine type, the wrong Social Security or serial number, or something else, it just writes back to the insurance company and says, ‘Could you please correct this certificate?’”  

Conversely, agentic AI is not best suited for “if-then” processes that are hard-coded, Verhelle said. For example, many CRM systems apply if-then coding logic that can be automated. 

“This is not what AI does best,” he said. 

Check out our exclusive industry data here.