![](https://www.constellationr.com/sites/default/files/styles/medium/public/supernova-photos/nicolas-moch-seb-2.jpg?itok=qVHYFvFF)
Nicolas Moch
Head of Info, Architecture & Strategy, SEB
Artificial Intelligence and Augmented Humanity
SEB is a leading Nordic financial services group, guided by a strong belief that entrepreneurial minds and innovative companies are key in creating a better world. In Sweden and the Baltic countries, SEB offers financial advice and a wide range of financial services. In Denmark, Finland, Norway and Germany, the bank’s operations have a strong focus on corporate and investment banking based on a full-service offering to corporate and institutional clients. As of 2015, SEB served 2,300 large corporations, 700 financial institutions, 257,000 corporate home bank customers, and 1.3 million private home bank customers.
SEB strongly believes that customer loyalty drives long term profitability and that a strong brand experience builds satisfied customers. Therefore, living up to our vision of providing world-class service to our customers is crucial for long term success.
In implementing Amelia, we sought to resolve inconsistencies in service; for example, some agents were using old processes to reset passwords or using old security questions the customer had forgotten. In addition, SEB had an issue with agents using long processes for simple tasks, creating a pain point for customers; a password reset took 20 minutes. Finally, we also were looking to remove inconsistencies from customer interactions, such as misinformation or lack of compliance with standards.
In SEB’s efforts to provide best-in-class customer service, we wanted to provide 24/7 access on multiple channels, without requiring overnight shifts or offshoring. While phone contact is the main channel of communication for external customer service, SEB is active on various social media channels and online via our website.
We began with an internal pilot, using Amelia in SEB’s service desk. To select requests to be automated, we looked at the most common & simplest-to-repeat tasks. Initially, 3 customer queries met these criteria: providing information on becoming a customer, ordering an electronic ID, and explaining how to do cross-border payments. When needed, Amelia redirects customers to the correct human agent. After the successful internal pilot, SEB launched Amelia for external customer service for over a million customers, making it more convenient to submit queries online & resolve issues faster. With Amelia, SEB gives an immediate response & ensures call volume for human agents is focused on high value support areas. Currently, Amelia handles customer queries like password resets, assistance with credit & debit cards, & banking location services. Her abilities are now expanding to complex tasks like ID verification after an account-sensitive question is asked.
Amelia’s intent recognition abilities allow her to respond faster to online requests & elevated the quality of the online banking experience. She has also shown measureable improvement internally in SEB’s service desk, improving typical resolution time for a password reset to 4 minutes, from 20 minutes.
Amelia has been live in her customer facing role since December 2016. Implementation was an iterative learning process, & she became a new channel customers can use to interact with SEB. In this role, Amelia has also reached a 90% accuracy rate in understanding & completing tasks, & correctly handing off to human agents those queries which she has not been trained on. She has improved service consistency, offers 24/7 availability & provides multi-channel delivery. Employee satisfaction has increased as agents now focus on higher value tasks. Agents enjoy learning & working with cutting edge technology, which is positive for SEB as an employer as people want to work for an innovative organization.
In addition, we have seen increased shareholder value as SEB has a competitive advantage over other banks through more organized & efficient customer service. Amelia has increased revenue by enhancing customer journeys, leading to more cross-sells & higher conversion & customer retention rates.
There were several metrics used to judge the success of Amelia. First, she is assessed on her ability to detect customer intent, understanding exactly what information the customer is seeking. Another key metric was her response accuracy, which is her ability to answer customer queries correctly. Finally, Amelia was measured on the number of customer queries addressed per day.
After implementing Amelia, her response accuracy reached 90% on the content that she is trained on and is continually improving as she addresses 300 customer queries a day. The rate at which she has been able to detect the underlying intent of a customer’s query has also increased on par with her new training and reached 85% during the same time.
IPsoft’s Amelia is the most human, comprehensive AI platform on the market. Amelia connects conversations to data and processes in order to provide a personalized service to every customer at scale. Amelia’s unique skills set includes conversational intelligence, advanced analytics, smart workflow, experience management, and supervised automated learning. Her roles span different functions for customer service and advisory support, to IT service desk and procurement.
SEB is faced with major disruption as the banking industry pivots towards a service-based model. Customers are increasingly fluid, both in channels used to communicate and ability to switch banks. SEB saw early on that cognitive technologies would transform the workforce and wanted to learn quickly as an early innovator at the forefront of disruption. Amelia’s initial implementation was a learning project conducted with active customers to shape the future of digital services. SEB wanted to see how customers reacted to Amelia, her effect on their behaviors, and how well she aided them. Customers were invited to try Amelia, with the disclaimer “Our chat is new and under development.” After interacting, customers gave feedback to be incorporated into Amelia’s functionality.
A challenge for any automation project is fear of staff reduction. However, SEB employees know that Amelia is a competitive advantage. She removes tedious, repetitive tasks from employees’ workflow, leaving more time for meaningful dialogue and relationship with clients. Employee performance is assessed on quality rather than productivity, as human agents now deal with complex, longer cases.
Amelia is an additional channel to reach new and existing customers- always available and infinitely scalable. SEB can interact with customers who would not walk into a bank branch or pick up the phone to speak to a human agent.
Just prior to launch, Amelia was answering many of the test queries incorrectly, and we were worried that we would be unable to meet our target date. Five days later, she had been fixed and was launched successfully. Amelia’s ability to learn quickly was mind-blowing, unlike a traditional IT project that goes through a succession of releases taking several months. Instead, this is a transformational business opportunity, allowing for a highly flexible and iterative way of working.