Hi Amit,

Thanks for responding.  We are using Customer Service 7.31 on Pega Platform 7.3.1.

Email triage will be used in a number of different methods, mostly using Robotics to extract PDF contents or NLP to extract entities on structured emails (i.e. emails that resemble forms with entities listed down a page).  The reason for using NLP in that case is we could have numerous number of emails with entities in different order.  I believe NLP is more reliable in these cases, do you agree?

One of our use cases in this respect is receiving an email with a number of entities which describes an insurance policy mutation.  The email has no other content other that policy number, type of mutation and some other required attributes.  The email is generated by an intermediator's website after a client requests a change of policy from their portal.  We would like to extract these attributes and call an external systems API to update the policy then respond back to the client submitting the policy change to the portal when successful.

We would also like a interaction created to give CSR's visibility of the email if the client may call the insurance company with any inquiry.