Planet42 had to tackle the automation of their document intake processes: the IDs, driver’s licenses, and bank statements that applicants needed to submit.
This strategic pivot towards document automation came with two clear challenges: document validation, and document understanding.
Each is critical, as automating without validating and understanding opens the door to automated fraud (a costly loss when the assets are entire vehicles).
But they're also individually mammoth tasks, and with existing processes manually driven, the team had little data to leverage and get started.
The company acknowledged early on that on-the-ground validation teams wouldn’t be able to keep pace with the growth in applications.
To maintain control and best performance, the Planet42 data science team opted to leverage Amazon Textract to build their own workflows — and Resistant AI to protect them.
Partnering with Resistant AI took one of those mammoth tasks off the table. Automating document validation allowed the data science team to focus their efforts while relying on Resistant AI’s expertise to detect signs of tampering in every document.
Resistant AI’s document classification data was also an invaluable component, jump-starting Planet42's technical knowledge to enable important automation design decisions early on (such as determining which bank statements to validate).
Resistant AI gave Planet42 a much better understanding of what their document set actually looked like, and where to focus parsing efforts. With evidence-based decisions, investor confidence grew knowing that new automations were well protected.
By helping turn a slow, manual application process into a fully automated one, it also allowed them to better securitize their debt.