Page tree
Skip to end of metadata
Go to start of metadata

Using Ephesoft Intelligent OCR

While there are many possible configurations and variations that utilize Ephesoft functions, using Ephesoft generally involves four basic steps:

A four-step overview of the process, 1. Add File for OCR, 2. Send to Ephesoft, 3. Receive Results, 4. Validate Results

This topic describes how to use Ephesoft's intelligent OCR capabilities to manage document conversions from PDFs into useful record data. 

Create OCR File Records

To be sent to Ephesoft, files must first be added to the Intelligent OCR Files table. There are many ways to achieve this, and the right solution for you will depend on your other workflows and configurations. For example, you might use a rule to take any files attached to incoming emails and convert them to Intelligent OCR Files records. In other cases, you might add an action button in another table that converts each attached file into an Intelligent OCR Files record.

Whatever method you choose, you must create a new record with an attached file in the Intelligent OCR Files table in order to begin the OCR process. When a new record is created in the Intelligent OCR Files table, its status is set to Draft.

Send Files to Ephesoft

After a record has been created and a file has been added to the Attached File field, the Intelligent OCR with Web Services action becomes available. You can choose to run this action manually, when a user clicks Extract Data in the Intelligent OCR Files record, or automatically, with a rule running the action when new records are created or a rule running the action on all new records as a batch at specified time intervals.

When you run the Intelligent OCR with Web Services action, the file is sent to Ephesoft and the record status is set to Queued. Ephesoft analyzes the file to identify the document classification and associated batch class and then extract the corresponding metadata.

Receive Results from Ephesoft

The system automatically checks for the data periodically until Ephesoft returns a result. You can also click the Check Status button in the Intelligent OCR Files record to manually pull results. Ephesoft returns the resulting data and confidence scores for each item when its analysis is complete.  Agiloft populates these results in the Intelligent OCR Files record according to your field mapping, storing both the field values and their confidence scores.

When the results are populated, the record status is set to Complete. You might want to configure a rule to check the confidence scores and set the status to something else if a certain threshold isn't met. For example, you might add a status called Ready for Validation and change the record to that status so you know it needs to be reviewed.

Validate Results Manually

If the confidence scores are low, you can simply open the original file and check the data in the file against the values populated in the fields. When the value doesn't match, simply edit the field and update the value. You might want to configure a way to highlight fields that fell below the confidence score.

Note that if fields were left empty, but the data was included in the original file, you should work with an implementer to check the Ephesoft setup and extraction rules.

Use Ephesoft Results

The same way you brought files into the Intelligent OCR Files table, it is often easiest to use the data by copying or converting it back out of the table when the record's status is Complete. How you accomplish this will depend on the type of documents you are analyzing and the processes you use it for, but you should consider what you want to happen when the status is Complete and configure rules, actions, and any other necessary entities accordingly.