Quadrant of Correctness – Enhancing Datacap and STP
Straight through processing is the ideal of any Datacap deployment. Despite how difficult that ideal may be to achieve, Datacap leverages multiple scanning engines and employs many strategies to improve touchless processing. Even with those tools, how can you guide your capture processes to achieve a greater throughput?
Allow us to introduce the Quadrant of Correctness. This is a concept conceived of by our very own Tom Stuart, who was the VP of Solutions at Datacap prior to IBM’s acquisition of it. By understanding how to classify the correctness of your document capture, you can then focus on making improvements.
When you capture a document, each of the fields will be classified into one of the quadrants below. The ideal is that all fields in the document will be captured correctly, and for nothing to be flagged.
Quadrant 1 – Here a field has been captured properly and without being flagged for verification, is automatically passed to another system. This could be simply to store it, or possibly to use in another automation process. This represents a field that has been processed correctly without any human intervention.
Quadrant 2 – Here a field has been found to require verification from a user and that it was flagged by the system to obtain a human review. While it does require human intervention, the capture process has worked properly.
Quadrant 3 – The field here has been captured without any errors but for some reason the system has suggested that it requires human-verification prior to final processing. The capture of the field worked properly, but being unable to be sure of this, the operator must look at it for assurance that it is correct. This is not only time-consuming, but if operators see too high a percentage of these, they can miss the few that are actual errors. After all, marking everything is the equivalent of marking nothing.
Quadrant 4 – In this last quadrant, the field was not captured correctly but the system did not identify any issues. This quadrant is the area of greatest concern as it has taken an incorrect value and moved it forward to a system of record, or other automation process. This is rare, but it happens enough that they gave it a name, a substitution error.
The Datacap validation rules play a significant role in helping to alleviate errors that occur in quadrants four and three – the two most erroneous areas that require the most attention. By focusing your efforts to in these two quadrants you are moving issues from quadrant four to quadrant two, and from quadrant three to quadrant 1 (see image below).
These validation rules help take issues from both quadrant three and four and move them up two levels. Fully leveraging these validation rules can offer a significant enhancement to any Datacap deployment. Truly, it is the rules-based engine that powers Datacap that gives it the strength to be the industry leader it is. While touchless processing is the ideal, and Datacap’s rules-based engine and validation rules help advance fields towards quadrant one, the power of the platform rests in the hands of people who know how to configure them.
Looking beyond the core Datacap functionality, the Cognitive Correction Module that MagicLamp created (and is available from IBM’s PartnerWorld) helps to further STP. This module learns the common corrections made by operators for common fields and automatically replaces the values. This learning and automatic adjustment help improve the number of fields that do not require verification. It also helps quadrant 3 items by looking at fixed but known values, like employee names, part numbers, addresses, anything where we can compare the captured result with a previously encountered value within their organization.
In addition to the validation rules that are part of Datacap, this enhancement allows organizations to move fields from quadrants two through four to quadrant one.
While many people seek straight through processing, it is of utmost important to not measure a project simply by this metric. Doing so fails to recognize the immense value Datacap provides by eliminating vast amounts of time and effort spent entering data.
For some Straight Talk about STP, check out this other article by Tom.
Attaining high levels of STP is dependent on a lot of variables not mentioned in this article. But, by leveraging the quadrant of correctness to understand where to focus your attention, you can optimize Datacap to further your touchless processing goals.