SEB is a leading Nordic retail bank, headquartered in Stockholm, Sweden. It has more than 15,000 employees serving more than 4 million customers across Europe.
In 2016, SEB had complex processes for resolving employee IT issues; simple requests took longer than necessary and the bank wanted to find a way to shorten mean-time-to-resolution. In addition to improving the efficiency of its support processes, the company wanted to free its experienced employees from high-volume (but low-level) tasks so they would have time to address more complex employee needs. SEB wanted to implement an alternative channel where it could provide customer-friendly support.
The bank initially implemented IPsoft’s Amelia, renamed Aida, for its internal IT helpdesk, handling password resets, network connectivity and firewall support, ordering supplies, booking meetings and troubleshooting business applications such as Skype.
Employees communicated with Aida in natural language via a chat interface, and she was trained to understand users’ intent for each request. When Aida was unable to handle a query, she escalated to a human who could view the chat history and pick up the request without interruption.
Following success of the initial internal deployment, SEB also introduced Aida as a Swedish speaking, external-facing chat agent in December 2016 to deliver scalable 24/7 customer support. She facilitated a variety of external customer service requests including guiding customers through ordering replacement credit cards, booking meetings with branch staff, and providing general account and branch office information.
Today, Aida provides IT support for 15,000 employees. Aida takes simple repetitive requests out of employees’ hands, allowing them to focus on higher-value tasks and more complex and engaging interactions. She especially helps meet demand during peak interaction times thanks to her 24/7/365 availability.
For customer-facing services, Aida handles queries from approximately 300 customers per day, performing tasks such as booking meetings, finding nearby services, and guiding users to open accounts.
Aida’s accuracy in recognizing intent is far superior to that of a static chatbot. She’s capable of properly determining intent during 93% of her conversations. If she can’t determine intent, she seamlessly escalates the conversation to a human employee. User response has been exceptional — 91% rate their experiences with Aida as “very good” or “good.”
Following the success of the original deployments, SEB has subsequently added two new Aida use cases. She’s now employed to assist SEB employees with human resources questions and she helps SEB customers research life insurance policies.