Many of our customers are using systems for automatic document classification and data extraction. ‘Kofax Transformation Modules’ (KTM) is one of these systems. These data capturing systems extract metadata out of the electronic images (these are the scanned pages of the documents, faxes or emails) and release the data and the document to business applications. KTM can be integrated as a module into the capturing solution Kofax Capture (see Stefan Blank’s blog article).
You can read more about classification and extraction with KTM in these older blog articles:
Document classification with Kofax Transformation Modules (KTM)
Kofax Transformation Modules – format locators and dynamic regular expressions – Part 1
Kofax Transformation Modules – format locators and dynamic regular expressions – Part 2
I will use the example of an insurance company and its CRM system, to explain how data extraction programs may achieve faster customer communication and hence achieve better customer satisfication.
In a process queue ‘mailroom –> application system’, you will try to get as many documents processed automatically , that is without any manual interaction (exception: scanning).
More details about process automation can be found here (sorry only in german):
Christian Böhnel’s blog article.
During our project at the insurance company (details below) we went through three stages of document process automation:
- manual: all incoming documents were processed manually
- half automatic: a subset of the incoming documents were processed automatically in the mailroom
- automatic: a subset of the incoming documents were processed automatically in the mailroom and in the application (CRM) system
We had to deal with three different sources for the incoming documents (order by frequency):
After the mailroom processing, all these documents have to be available as an attachment within the transactions of the CRM system. For that, you have to add the needed meta data to the documents. More and ‘better’ meta data will lead to a faster and target-oriented processing in the CRM system. Examples for meta data are: class of insurance, insurance number, document type, …
But first of all the incoming paper documents have to be digitized by scanners, the e-mails have to be converted into bitmaps and faxes are already digized by the fax servers.
Our customer is using the capturing solution Kofax Capture. After the digitisation of the documents, all meta data was entered manually by the staff of the ‘document service’ department. This was done by a Kofax validation tool.
As the documents are organized in batches, all documents of a batch had to be indexed with meta data, before they could be routed to the CRM system. If the meta data for only one document could not be entered at the moment, the whole batch was delayed.
At this stage of the project two ways of optimizing the throughput time came to mind:
- Automatic classification of documents and automatic extraction of the needed meta data
- breakup the fixed batch structure during the processing of the documents in the ‘document service’ department
For item 1. we installed the Kofax product KTM (Kofax Transformation Modules). KTM can be seamlessly integrated into the capturing solution Kofax Capture. Now the meta data could be extracted automatically from the documents. To improve the data quality the meta data was checked against the database of the CRM system – if possible.
Nonetheless all documents of a batch were reviewed by the ‘document service’ staff and the meta data was corrected or added. As we did the check against the CRM database, it was now possible to skip the valid documents in the validation tool. After some time for ‘confidence-building’, we enabled the skipping of valid documents.
Though valid documents werde skipped in validation, the staff was still able to select and check these documents if desired. At this stage the manul data entry was supported by the automatical extraction, but the fixed batch structure still prevented the quick routing of valid documents to the CRM system.
Now we implemented another KTM tool, ‘document routing’, that splits a batch into two independent batches after data extraction and before the manual validation step takes place (technical details are provided in the scripting at the end of this article):
- Batch A contains all valid documents of the original batch. These are the ‘database checked’ documents
- Batch B contains all other documents of the original batch
The documents in batch A can now be routed directly to the CRM system without any further manual interaction. Documents in batch B are manually validated by the staff as before.
At this stage the document process automation looks like this:
Document process automation for the whole process queue cannot be achieved by the capturing solution. ‘Document service’ can only provide all needed documents and their meta data in the best possible quality to the target CRM system. The application logic of this system has to determine, if data can be processed directly without any further manual interaction. If this is the case, we have document process automation for the whole queue from scanning to booking with as little manual work as possible .
KTM contains a tool called ‘Document Routing’ for routing documents from one Kofax Capture batch into a new batch directly after the data extraction step. Furthermore you can also define the next step in the Kofax Capture queue for the new batch (in our project: the routing to the CRM system).
This is an example coding in the KTM scripting language, which shows how to use the ‘Document Routing’ feature:
' Class script: Project
Private Sub Batch_Close(ByVal pXRootFolder As CASCADELib.CscXFolder, ByVal CloseMode As CASCADELib.CscBatchCloseMode)
If CloseMode=CscBatchCloseFinal And Project.ScriptExecutionMode=CscScriptModeServer then
Dim oXDocInfo As CASCADELib.CscXDocInfo
Dim i As Integer
Dim Batchname As String
'Name of the original batch
Batchname = pXRootFolder.XValues.ItemByName("AC_BATCH_NAME").Value
'loop over all documents of the original batch
For i = 0 To pXRootFolder.GetTotalDocumentCount - 1
Set oXDocInfo = pXRootFolder.GetDocInfoByGlobalIndex(i)
'Is this document valid?
If oXDocInfo.XDocument.Valid = True Then 'mark it for the new batch
'Name the new batch just as the original batch with attached '-New'
pXRootFolder.XValues.Set("KTM_DOCUMENTROUTING_BATCHNAME_ChildBatchNew",Batchname & "-New")
'next Kofax Capture queue for the new batch is Export