AI Mapped Fields
Disclaimer
While our document scanning and data extraction software leverages the powerful capabilities of Large Language Models (LLMs), it is important to note that LLMs are not infallible. There may be instances where the software produces inaccurate or incomplete data. As such, the responsibility for verifying the accuracy and completeness of any data collected through the software ultimately lies with the user.
By using this software, you acknowledge and accept that the final onus of ensuring that the extracted data meets your requirements and is correct is your responsibility. We recommend that all data be thoroughly reviewed before being used for any critical or decision-making purposes.
When mapping AI fields to regular Scan2x metadata fields, besides the regular mapping method, there is the below method which may be more beneficial to use when mapping metadata of multiple documents.
When an AI module is chosen, the following button will appear for the administrator.
Once the Manage Mapped Fields button is clicked, the administrator will be shown the following screen.
|
Metadata Field
|
Select the metadata field that you would like to map AI fields into. If no metadata has been created in the Main Metadata tab, this dropdown selection will be empty.
|
|
|
|
Use Job Template
|
Scan2x will use the job's document template and analyse it for AI fields.
|
|
|
|
Specify new file
|
Scan2x will use the document specified in the below file path.
|
|
|
|
Analyse Document
|
Initiates Scan2x to analyse the chosen document for AI fields.
|
|
|
|
Analysis Results
|
The AI fields analysed from the chosen document will appear in this tab.
|
|
|
|
Label Settings
|
The administrator will be able to filter the information that is picked up from this field.
|
|
|
|
Table Settings
|
The administrator will be able to enable further settings for AI fields which are tables.
|
|
|
|
Add Document
|
The administrator can add more documents to be analysed by the AI
|
|
|
|
Analysed Documents
|
The documents analysed by the AI will appear here.
|
|
|
|
Apply
|
Click Apply to save the changes made.
|
|
|
|
Cancel
|
Click Cancel to exit the Managed Mapped Fields tab.
|
|
|
|
New Metadata Field Name
|
Adding a new Metadata Field can be done from this function directly. The new metadata field will then be shown in the Main Metadata tab or Temporary Metadata tab once the 'Add Metadata Field' button is clicked.
|
|
|
How to map AI fields to Metadata Fields
The administrator has the option of analysing the Job's document Template or can specify a completely new file for the Scan2x AI to analyse.
Once the Analyse Document button is clicked, the administrator will get the following warning: "Please note that clicking analyse will use 1 or more Scan2x AI Credits each and every time it is pressed. We will store the results of each analysis for future use in the Scan2x Job Designer. Proceed with running analysis for this document?"
The administrator can choose not to be notified again or can continue the process by clicking OK. If the administrator does not want to continue the process, they can click CANCEL.
Once the OK button is clicked, Scan2x AI will then analyse the document and in the Analysis Results tab, the AI fields will appear showing their name in the Label column and the information data picked up from the document in the Value column.
From here, the administrator can choose to map specific AI fields to the metadata field specified in the top-left corner dropdown.
If the administrator requires it, they would be able to add and analyse more documents to map to the same 'InvoiceDate' metadata field.
This can be done by clicking the Add Document button in which an empty row will be created. The administrator then needs to tick Specify new file and click the folder icon to select the desired PDF.
Once the file is chosen, click the Analyse Document button and click OK for the warning pop-up (unless already set to do not disturb). The modal will then show the AI fields of the new document, however the AI fields list show the information:
•In green if there are AI labels that are common with the previous document (i.e., the Job Template in this case).
•In black if the metadata of this newly analysed document (i.e., Canon) does not have common AI labels.
The administrator can also choose to create a new Metadata Field, without needing to close the Manage Mapped Fields tab:
•First add Name of the new Metadata Field, click Add Metadata Field
•From the Metadata Field dropdown, Select the newly created Metadata Field
•Map the new field to a Label, and Apply changes.
The administrator can also adjust the selected/ticked AI label settings from the Label Settings tab.
As can be seen in the image below, the administrator can decide to parse the InvoiceDate AI field as Text, Date (system format), Date (UK format), or Date (US format).
Once selected, the administrator can then go a step further and select a FilterType and FilterCondition, such as the image below where the InvoiceDate will show a result if it contains the string ' 2023'.
For more information on the Table Settings tab, kindly see the Table Definition with AI tab.
Once the administrator has mapped the AI fields for the metadata field InvoiceDate accordingly, click Apply.
From the Metadata tab, the administrator will see the InvoiceDate metadata field as a Mapped Field with the AI fields InvoiceDate mapped to it.
Remember to click Apply Changes to save the job configuration changes.
Administrators can also click the button highlighted in red in the image below to search and select one or more Mapped Fields.
If administrators would like to add all the AI fields found from the Managed Mapped Fields tab automatically, click the Insert All Scan2x AI Cloud Services Fields button.
|