In this episode of Tech Solutions for CUs, host John San Filippo interviews Ronny Chapman about his new company, VisionFI. Chapman explains that VisionFI is not a standalone product but a component that plugs into other solutions like loan origination systems (LOS) or new account opening platforms, and even Microsoft Office.

VisionFI's core function is to use multimodal AI to read images, photos, and documents, taking unstructured data and bringing structure to it. Chapman provides an example of its original use case: taking a picture of a car's VIN through the windshield to extract the VIN, eliminating manual input. The technology has since evolved to process various documents such as tax returns, operating agreements, and auto loan packages.

A significant use case in auto lending involves members bringing paperwork from a car dealer to the credit union. VisionFI can process this paperwork, like retail installment contracts and buyer's orders, by simply taking a photo, converting what would be a 30-minute manual entry process into an automated input for the LOS. This effectively turns a dealer-financed loan into a direct loan or refinance for the credit union.

Chapman also discusses how VisionFI can be used for loan quality checks, aiming for 100% accuracy rather than sample checks. He emphasizes that VisionFI doesn't add judgment or decisioning but reads information faster than a human, allowing employees to focus on higher-value tasks.

Regarding pricing, VisionFI charges per 1 million "tokens," which represent computational thought and are a measure of AI consumption. Chapman estimates that humans process around 500 tokens per minute, while VisionFI achieves about 26,000 tokens per minute. He suggests that a human's token processing can cost $300-$700 per million tokens, while VisionFI could be $50-$60 per million tokens, demonstrating a significant ROI.

Chapman shares an aspirational use case where VisionFI could allow members to simply send documents like tax returns, pay stubs, or W2s, and the system would automatically convert and input the data into an application, eliminating the need for manual keying and enhancing member service. He also predicts a future where personal consumer tools, like Apple iCloud or Google accounts, will be more integrated with financial services for seamless document sharing and online banking logins.

Credit union executives interested in learning more can contact Ronny Chapman at rc@visionfi.ai or visit visionfi.ai. He welcomes discussions on potential use cases and opportunities to improve member experience.