Swiss Post Solutions

Efficient robots for PostFinance payment collections

Since autumn 2016, PostFinance and Swiss Post Solutions (SPS) have been testing how software robots can support the PostFinance payment collection process in a joint project.

If Mr Smith or Ms Jones had forgotten to pay their credit card bills, a complex process used to be triggered at PostFinance: a list of the details of defaulting customers was processed manually at regular intervals by a back office team at the bank. There was a multistage process to be considered and various systems had to be maintained.

Process automation with robotics

Swiss Post Solutions developed a solution for PostFinance using robotic process automation (RPA). A robot goes through the accounts of the customers in question at regular intervals – including outside office hours – to enable the efficient processing of payment collections. It recognizes data such as the IBAN number or credit card type and launches various back office applications to analyse available assets and check when previous credit card bills were settled. In the next step, the robot notifies employees of accounts with billable amounts. The robot therefore supports the process, almost in real time and taking all necessary business rules into account.

Better quality and compliance

“RPA is a key technology that enables companies in times of digital change to improve processing quality and reduce manual processing time,” says Jörg Vollmer, Head of Swiss Post Solutions. “Another advantage is that the old systems do not have to be reconfigured because the robot works on existing systems. “There is also a guarantee that each step is fully traceable and documented and that the data concerned remains in the company.”

Intelligent automation

Artificial intelligence can automatically process unstructured documents such as e-mail enquiries. First it determines the content of the message: for example, is the message a request for a quotation? The system then looks for key information such as the name of the sender and their customer number and combines it with existing internal customer data. Depending on the data available and the rules defined, the system can now continue working completely independently and send the requested quotation. If the system cannot manage without help, it forwards the request to an administrative assistant. If the same situation arises again, the system will recognize the case and deal with it independently. It is therefore learning continually.