Imagine waiting for a diagnosis, asking someone to marry you, or any other number of requests for decisions that could completely change your life and then waiting for weeks for an answer. With life savings and a lifetime worth of dreams hanging in the balance, these were the stakes for PPP loan applicants and the banks trying desperately to serve them.
The $349 billion Paycheck Protection Program (PPP) was quickly conceived and launched to help save small businesses battered by the pandemic’s effects. But what was meant to be small business lifelines became tidal waves of loan applications. In the first hour, on a Friday, Bank of America received 10,000 loan applications from small businesses desperate to stay open and pay their employees.
By Monday morning, the bank had received 177,000 applications—for a total of $32.6 billion in loans—or nearly 10% of the $349 billion allotted. The speed at which the PPP loans needed to be approved and processed was essential to the program.
Too big, but can’t fail
Banks simply could not keep up with the unmanageable amount of documents flooding in for every single applicant. Further, employees faced intense pressure to process applications on top of their current work quickly and with unprecedented demand when nearly every employee was working nearly exclusively from home. Because much of the information needed to process loans was inside emails and documents, it created enormous delays.
Employees spent much of their time collecting, extracting, and collating the data from disparate formats. At another global bank, an executive describes frustration-filled daily conference calls with hundreds of employees across departments, trying to find solutions to speed up the application processing.
That bank finally decided to entirely outsource the review process offshore, employing people to sift through and process loan applications manually, driving up the overhead with a fixed cost per loan processed. Given the significant and well-publicized process challenges, the question becomes, “What can we learn?”
The problem: the data extraction logjam
Freeing business data from assorted emails and documents is a high-touch, high-friction process. Loan applicants must submit forms, emails, and many supporting documents—utility receipts, payroll reports, and rent receipts—to meet eligibility requirements and obtain loan forgiveness.
While some banks have adopted automation solutions such as robotic process automation (RPA) to speed up workflows and improve processing times, most PPP applications contain a lot of unstructured data, which RPA can’t use until it has been structured.
As a result, nearly 80% of an employee’s time is spent gathering the data, with only 20% spent analyzing and making the decisions. Besides prolonged turnaround time, manual processing drives up operational costs, increases errors, and might even result in data omission, creating compliance headaches and regulatory exposure.
The solution: next-generation technology and a new approach
By automating unstructured data extraction from emails and documents, and turning it into structured data, Alkymi handles what RPA can’t do well. Handing business users— loan officers—a technology solution would reduce the work from hours to minutes and increases crucial decision-making.
When intelligent workflows using machine learning can handle data extraction, loan experts only have to review anomalies or low-confidence data—essentially teaching the machines what to look for next time—and the loan is processed in a fraction of the time.
Just imagine the value that can be created when these operational efficiencies are realized. No more conference calls with hundreds of people. Customer experience can be real-time digital interactions. Employees are relieved of manual data processing and banks can make real-time decisions. With so much weighing on an answer—businesses, jobs, lives —every hour saved can make all the difference.
To learn more about Alkymi’s approach to data extraction, schedule a demo.