ICE vs OCR
Most invoice capture providers use one of two methods to capture invoice details:
- Straight single pass manual keying where the accuracy of the capture is fully dependent on a single human being’s ability to accurately key invoice information from an image.
- Optical Character Recognition (OCR) where a computer based program makes a single pass to extract invoice information accurately and relies on a human being to validate any fields the OCR program cannot extract.
While OCR is generally perceived to be the more accurate and forward thinking approach to invoice capture, OCR is not a perfect technology as OCR still results in a significant number of invoices that have to be validated or corrected by a human being post the OCR process. That is where CTI’s ICE process differentiates itself from the providers who just manually key or implement some form of OCR for invoice capture.
The significant advantage of ICE is that the same data on a single invoice is extracted not once but twice using different capture intelligence levels. Then the technology of ICE programmatically compares the dual capture to verify if the fields extracted by both processes are identical. Then only if both capture processes result in a difference does the ICE process route as an exception to make a final determination as to the accuracy of the field. CTI handles all the invoice exception cleanup rather than pushing the cleanup to the customer like traditional OCR. The ICE dual extraction process results in less exception cleanup since the invoice data is being extracted twice on the same invoice and compared vs OCR where the data is extracted only once before sending to a human being for review. As a result, the ICE process can deliver minimum accuracy levels on invoice capture of 99.95% which is significantly higher than either OCR or just manual single pass capture. Additionally being there is less data validation needed on the invoice capture using ICE, invoice turnaround time can be under 24 hours from receipt.
ICE Benefits vs OCR
- Minimum 99.95% capture accuracy
- 24 hour turnaround from invoice receipt to visibility in your ERP system
- No OCR type exception cleanup needed by your AP department
- No OCR templates that need to be built by invoice format
- No scanning by your AP team for paper based invoices
- Rules engine to handle automated exception handling
- Notification to suppliers or end users of invoice exceptions
- Software as a Service model means no software to buy or support
- Handles all invoice types (paper based, emailed and faxed)
- Handles all image types (PDF, TIFF, JPG, Excel, Word, etc)
- Lower cost than OCR