Search automation in Portran (Ipiranga)

Issue:
Portran from Ipiranga provides carriers with a dashboard that allows searches by base (CNPJ) and start/end date, enabling them to identify which freight entries already have a link, receipt, and have been paid. However, there are some problems. Searches are limited to one base; searches are limited to periods of up to one month; and there is no notification of old freight entries that have been paid (requiring periodic validation).

Therefore, the finance department needs to check Portran daily (and perform monthly audits) for all past payments that may or may not have been paid, and this task of searching all databases for every month takes around 4-6 hours.

Solution:
Automate daily searches in Portran using RPA, maintaining an up-to-date proprietary database with information, making queries faster and more personalized, including sending notifications via WhatsApp when a specific old transaction has been linked, received a receipt, or received a payment.

Thus, I set up the following structure: (1) the bot enters Portran and, for each database of interest, searches month by month for entries; (2) each entry found is saved in a local .xls file; (3) a second bot monitors the entries folder, sending them to the main server when it finds something; (4) the server receives this information and saves it in the database; (5) the server also validates if, in the case of an entry that already existed, there is a change in status (receipt, link, and payment); (6) if there is, our AI (Beto) sends a message to finance with the entry number and what happened.