OCR capability / Using Tulip to identify batch numbers from camera captures

We’re looking to explore the possibility of elimianating a manual/visual verification of paper records by the QA personnel on the warehouse. Basically before releasing the batch for courier/shipment the QA personnel will manually verify if the batch number matches in the picklist and the labels generated by SAP.

In Tulip we’re introducing a camera capture widget to upload the picklist and label photos into the record but we wanted to take it a step further to see if Tulip can automatically recognize the batch number from the photos and verify that the numbers match in the two photos.

Below is the use case that we would like to explore Tulip’s capability on -

Step 1. Operator takes a picture of the entire Pick List which will have a bunch of information but the information which we need Tulip to read and extract is the Vendor Batch highlighted in red below.

Step 2. Operator then takes another pic of the entire secondary label where we need Tulip to be able to capture the Batch number highlighted in red below.


  1. We are looking to see if Tulip could compare the batch number captured from both the pictures and ensure it is the same.

Hi Avi.

There are a few ways to address this.
The best and easiest will be with Tulip Frontline Copilot to extract text from images (see KB article here).

This will allow you to take a picture of the documents, extract the Batch Number (the Queries can be “What is the Batch no?” and save the results to a variable.

Then repeat it with the Secondary label and compare the two with by checking if one variable Contains the other.

To use copilot you will need to join the beta program. If you need it for LTS environment which doesn’t have the co-pilot features yet. We can accomplish the same by using Connector Function to AWS Textract service (as described in this KB article).

I’ll be happy to get on a call to help you out with this.

1 Like