As an initial step we would like to give a brief overview of our terminology and wording. All apps offered are based on a single technical platform: our AI-Agents. As you may have already guessed AI is the acronym for Artificial Intelligence. Therefore we subsequently will often use the term AI-Agents.

This Ticket Enrichment App is also based on our AI-Agents‘ platform. Artificial Intelligence mainly works on the bases of so called recognition patterns. In order to create recognition patterns it is necessary that they are taught in once. In practice, this means that you have to teach in each document that you want to have recognized later on once.

For each taught-in document a four digit alphanumeric code respectively an AT2X, the so called DocTypeId, is automatically generated. The pattern is then saved under this DocTypeId by our AI-Agents within the AI. Several DocTypeIds form the so-called DocTypeSet, a collection of many documents. 

In order to avoid a separate teach-in of every single document with only minor deviations an algorithm makes sure that similarities to already taught-in documents (DocTypeIds) are recognized.

Thus, for example, it is sufficient to learn-in an employment contract once. You don’t need to teach it in separately for each employee just because of different employee names.


Instructions for how to teach-in documents

step-by-step guide


Step 1: Get your AI-Agent up and running!

Create a new ticket in your ticket system just the way you normally do.


Step 2: Feed your AI-Agent!

Upload a document (.pdf) to the created ticket that you want to be taught-in and in future be recognized by the Ticket Enrichment App.


As for the teach-in process it is fundamental that your document has a high quality in terms of its resolution. The better the quality of your documents is during the teach-in process the better will be their recognition in the future. Your documents should therefore have a resolution of min. 300 dpi or even higher.


It is even better to use generically created documents, e.g. by saving the document directly from MS Word as a .pdf file and uploading it. By doing so you avoid a loss in quality that normally occurs from printing and re-scanning a document.


Important: Only one document (comprising all pages) per ticket. This ensures that Ticket Enrichment App understands the size/extent the document actually has. Please make sure that the order of pages is correct when learning-in multi-page documents. Also, please never learn-in several documents in one ticket!


Step 3: Tell your AI-Agent what you expect of him! Customize Field Subject

We have pre-entered default settings as standard so that you can begin with your work right away. You can easily adapt/customize these defaults setting according to your individual requirements and needs.


The default-setting definition of the field “Subject” is always:


 


In dependence on the DocTypeIds you may also add document-internal information to the field “Subject” within the ticket. 


The following parameters apply:


Name

Parameter

Example Value

EmployeeGroup

(1)

0499

EmployeeNo

(2)

03173739

LastName

(3)

Doe

FirstName

(4)

John

DocDescription

(5)

Employment contract

OriginalSubject

(99)

SCAN PRAT00020288

DocOCR Text

(80)

Full OCR Text

Subject.DefinitionString = (1)/(2) | (3) (4) | (5) | (99)

If you want to make these changes please insert this Subject.DefinitionString in the field “Subject” accordingly.


 


Step 4: Give further instructions to your AI-Agent! Customize Field Body

As default settings no changes are made in the field “Body”. Here as well it is possible to adapt/customize these defaults setting according to your requirements.


Body.DefinitionString = (1)/(2) | (3) (4) | (5) | (80)


If you want to make these changes please insert this Body.DefinitionString in the field “Body” accordingly.


Step 5: Your AI-Agent can achieve more! Customize Field Customs

As default settings no changes are made in the field “Custom”. Here as well it is possible to adapt/customize these defaults setting according to your requirements. Please enter the desired value in the affected Custom-field. 


For text fields you can also insert the above-mentioned parameters which will be replaced afterwards!


After reading the ticket the AI recognizes existing “Custom” fields and takes over their pre-settings for this DocTypeId accordingly.


Step 6: Take your AI-Agent towards the finishing line! Customize FileName:

As a default setting  the definition of the  filename is always:


PN0499_03173739_Doe_John_Employment_Contract.pdf


FileName.DefinitionString = PN(1)_(2)_(3)_(4)_(5).pdf


 


If you want to make this change please name the file according to the FileName.DefinitionString before you upload it.


Step 7: Show your AI-Agent where to go! Transaction

In order that your AI-Agent understands the mission you have for it you have to transfer a TransactionsCode to it. This is done as follows:


Order/Process

TransactionsCode

Learn-in a new document

*** AI.Learn.Doc.Create ***

Change a document e.g. DocTypeId: AT2X

*** AI.Learn.Doc.Update.AT2X ***

Delete a document e.g. AT2X

*** AI.Learn.Doc.Delete.AT2X ***

Please enter the TransactionsCode in the first place of the “Subject” field of the created ticket. For example, you may complement the Subject” field

0499/03173739 | Doe John | Employment contract 

  with the TransaktionsCode: *** AI.Learn.Doc.Create *** as follows:

*** AI.Learn.Doc.Create *** 0499/03173739 | Doe John | Employment contract

 


Now your AI-Agent knows that it hast to learn-in a new document. 


Step 8: Send your AI-Agent to a journey!

You can now assign the ticket to the agent in the ticket system that you have set up for this app or A-Agents. In case you have several AI-Agents in use, one of them is sufficient. AI-Agents communicate and share their knowledge!


Step 9: Your AI Agent crosses the finish!

Within the next minutes your AI-Agent will process the ticket and learn-in the document accordingly. You can witness the result directly in the corresponding ticket!


Step 10: That’s it!

Now you can proceed to Step 1 with the next document.




This is just a small extract from the wide range of support our AI-Agents can provide your business with. Typically, an AI-Agent can process about 80% of all clerical routine tasks. As for the ticket enrichment you can also chose from a variety of functions. If you need additional information, please contact us

For example:

  • Assignment & Workflows according to document DocTypeId & employee
  • Reading.out of textual segments
  • Reading-out of information from documents and handing over to 3rd Party System like SAP HCM, Oracle, etc.
  • Processing of BarCodes
  • Provision of image sections from documents
  • API connection to HCM systems for validation of employees
  • API connection to HCM systems for transfer of additional information from documents 
  • API connection to HCM systems for transfer of documents & metadata