A new colleague who works quietly - even when everyone’s gone home

Since her new colleague started sitting next to her and taking responsibility for some of her work, Petra Maurer can attend to about twice as many business customers and private clients than she could before. “He is a quiet colleague, but very efficient,” says the real estate adviser at a regional bank with a smile. “And he works without ever grumbling - even at night.”

The new colleague she is praising is not a human being, but an algorithm. This clever set of calculation rules works quietly but powerfully at the heart of special software installed on Maurer's workstation. It relieves the real-estate adviser of humdrum aspects of communicating with customers - for example, by collecting information and suggesting e-mails wordings. 
 
Artificial intelligence (AI) in process automation is set to transform bank operating models. So-called intelligent automation will change the day-to-day routine of bank staff and their clients. Thanks to AI, many processes that previously required a human brain will soon be automated. 
 
And automated reliably  – processing of texts, understanding of handwritten entries or comparing of legitimation documents will soon take pace largely without human intervention. Ms. Maurer says she “often only looks only briefly” at the e-mails written by her algorithmic colleague. “It is really amazing how appropriate the e-mail wording proposals are and how well the algorithm formulates them. You really can’t tell that they were done by machine.”
 
Intelligent automation promises more cost efficiency
 
For banks, the technology means that significant improvements in sales and cost efficiency are coming into view. First experiences with AI-supported systems suggest that processing times can be reduced by up to 80 percent. This means customers could soon enjoy much shorter waiting times and unprecedented transparency about the processing stages of their requests. 
 
In Germany, more and more banks are beginning to switch to in-house automation of standard processes or outsourcing them to IT service providers. In five years at the latest, automation of these areas will be standard – and in the middle of the coming decade, the next step up will likely become the new normal: more advanced AI-supported methods such as natural language processing will carry out context-based processing of complex customer-application data. 
 
Banks must gradually prepare for AI deployment 
 
On the road towards automation, banks will have to prepare in two ways: they will have to standardize and digitize processes and use of existing network and core banking standards; and they will have to use automation solutions already available in the core banking systems.
 
Automation improves processes and the customer experience. Take the example of financing the construction of a new building, says Maurer. When she used to prepare applications by hand, she always had to check which attachments were necessary. Now, the algorithm does all this. "To be honest, he can do it even better than I can, he's never tired and doesn't make any careless mistakes,” she says. And if the algorithm is unsure, it will ask a human for help. 
 
As many careless mistakes and symptoms of fatigue are now a thing of the past, process quality has improved. For bankers like Petra Maurer, this means they have more time to concentrate on areas in which human skills remain essential - and will remain so for the foreseeable future: advising customers, being creative, thinking holistically, and communicating efficiently. 
 
"We now have more time to talk to a customers on the phone or to meet them in person," says Maurer. She says she can now get a better feeling for her customers’ wished and circumstances. In this bank, humans and algorithms work well side by side. This good coexistence has taken the fear that the two sides could work against each other. As Petra Maurer now knows: “The algorithm performs routine tasks that had previously annoyed more than delighted us.”