AI in OCR engines
In today’s world, the emergence of AI and cloud computing presents a remarkable opportunity. Deep learning AI has made OCR engines much more affordable and flexible. No more templates and zones. And the accessibility of cloud computing minimizes IT involvement. Consider this scenario:
- Place your lockbox file with check images, exactly as received from the bank, in a directory on your existing AR system
- A small cloud interface agent picks up the file and moves it to the cloud
- In the cloud, the lockbox file is parsed, check images are passed through the OCR engine and remittance detail is extracted.
- Still in the cloud, payments and their invoice applications are formatted into a file that your own ERP can import directly into its AR system
- This import file is transferred from the cloud back to your AR system and imported
- Your cash appliers now have all the checks from the lockbox already applied with the customer remittance advice supplied. They can work and post those checks in your existing AR system.
This approach means that power and productivity of OCR and industry-leading cash application techniques are now readily available to virtually any business:
- There are no additional hardware requirements
- There is minimal IT involvement
- There is no complicated software to install for OCR or a “bolt-on” front end to your AR system
- There’s no training requirement, because everything is delivered to your existing system
The benefits and economics of this scenario are compelling. For a nominal setup fee and an affordable monthly subscription, typically 70 to 80% of your lockbox check applications can be done automatically, freeing up headcount for more productive redeployment. And if we get more sophisticated and upload additional data, such as open item and customer information, and setup a few business rules, the automatic application rate can approach 95%.