Openlane, Inc.: Automating Unstructured Data with Unprecedented Accuracy and Speed
A subsidiary of KAR Global, Openlane, Inc. is the online vehicle remarketing and portfolio management pioneer, delivering intelligent innovations and trusted technology solutions preferred by most OEMs, automotive finance companies, and fleets with 20 years of industry leadership and a proven track record of delivering the latest technology at scale.
Based in Redwood City, CA, Openline operates more than 40 private-label remarketing marketplace sites and mobile apps in the U.S. and Canada, making it, by far, the largest provider of vehicle portfolio management and remarketing services in the industry.
Core Challenge: Large Volumes of Complex Documents Requiring Rapid Processing
Janet Narduzzi, Director of Finance Operations for Openlane, explains: “Openlane sells vehicles online with a value proposition of fast, seamless, and painless experience. Our goal is to remove friction from the purchasing process and make it easy for our customers to purchase cars in a wholesale market.”
The biggest challenge Openlane faced in providing a seamless customer experience centered on processing a large volume of vehicle titles in a timely manner. Openlane was tasked with processing 2,500 to 3,000 titles per day. With technology failures behind them, Openlane resorted to manually processing each title, a process that was both time-consuming and prone to human error.
There were several reasons this presented challenges for the company. First, vehicle title formats vary from state to state, making it difficult to identify the key information needed from each individual title. Furthermore, each state’s title format includes anti-fraud technology designed to keep fraudsters from duplicating titles successfully. While this technology is good for fraud detection, it presents significant challenges for digitization technologies such as OCR because it creates additional “noise”, which the document understanding technology must filter out. Openlane also receives titles in multiple ways, including physical titles sent via postal service and PDFs with hundreds of titles attached to emails. This requires an agile technology that can process information from disparate formats.
Accuracy in title processing is key to providing a good customer experience at Openlane since accurately identifying the vehicle identification number (VIN) of each vehicle ensures that all workflows and customer paperwork are handled correctly. However, due to the reasons outlined above, automating this process without compromising on accuracy seemed impossible with traditional methods.
Before partnering with Instabase to tackle the issue, Openlane had tried various other technology solutions with less than satisfactory results. Although Openlane is a digital-first company, the highly variable, unstructured data in titles was a pivotal constraint in their automation journey.
Stephen Herlick, Director of ERP for Openlane, noted: “We even bought a solution, but ultimately, the accuracy rate stayed at 20-30%, which was simply not acceptable.”
Openlane needed a different approach.
Solution: Openlane Partners with Instabase to Automate Unstructured Data at Scale
Openlane was determined to find a solution that enabled both accuracy and speed, which would, in turn, give their customers the superior experience Openlane wanted to provide.
Herlick identified some must-have attributes for a technology vendor in his search for a solution. Those attributes included:
- The ability to bring the best-in-class technologies together to provide consistently high accuracy and flexible, fast, and future-proof solutions,
- The ability to offer a turnkey solution so Openlane would not have to buy or host the infrastructure or manually scale for throughput,
- Seamless integration with Openlane’s existing technology stack, including SAP & AWS.
Herlick’s search led him to Instabase, the Automation Platform for Unstructured Data that uses cutting-edge deep learning technology and an innovative modular approach to rapidly create customized automation solutions to suit any document use case. The Instabase platform can be seamlessly integrated into any technology stack, and solutions built on the platform can scale to thousands of documents per hour without manual IT intervention.
The Instabase team built a custom solution for Openlane that automatically extracts the VIN information from title documents with 95% accuracy, which is then passed to Openlane’s downstream system.
Results: Instabase Brings Speed and Accuracy to Openlane
Right out of the gate, Instabase wowed Openlane with its accuracy and speed in title processing. Herlick related: “We got an 80-85% accuracy rate right away, and once Instabase began to finetune the solution, the accuracy continued to climb higher and higher. The solution now consistently hits the 95% accuracy mark.”
Instabase is not only accurate but speedy as well. The custom-built solution dramatically reduces the time and effort required to extract VINs from title packages. What previously took end users multiple hours filled with manual labor now takes a matter of minutes, thanks to Instabase.
Narduzzi said, “Since we deployed our solution with Instabase, I cannot think of a day that has gone by when we have not processed and shipped out any title that needs to be shipped out on the same day we received it.” Herlick added, “The level of technical expertise and partnership Instabase provided was critical to our success.”
Looking to the Future
Openlane’s digital transformation journey with Instabase is ripe for multiple use cases Currently, Openlane and Instabase are working together to streamline check processing, a task that sometimes involves as many as 170,000 checks per year. This is a critical process as clearance of the check dictates when a vehicle can be released to the customer.
Instabase built a solution that leverages cutting-edge deep learning technology to automate this process end-to-end. Narduzzi stated: “The solution configured with Instabase helps us extract information with unprecedented accuracy, execute on complex validations against the data in SAP, and run this process end-to-end in a fraction of the time.”
All of this is just the beginning on the road towards full digital transformation. Narduzzi added: “We’re looking at other areas in the organization and our sister companies that can be digitally transformed with Instabase and benefit from Instabase’s deep learning technology.”
Herlick couldn’t agree more, adding: “No matter if you have a lot of developers or none, you need solutions like Instabase, which is customizable to best suit the needs of your team. We will continue to tackle new use cases with Instabase as we see them not only as a technology vendor but as our long-term strategic partner.”
To learn more about Instabase’s unique approach to automating unstructured data with deep learning, download “The Ultimate Deep Learning Guide for Unstructured Data” today.