AI chatbot to offer customer care at Schibsted
Media Leaders | 23 October 2017
Chatbots have proven very useful in the e-commerce industry, and we hope our first, very own chatbot will do the same for our subscription media.
Altogether, Schibsted Norway Subscription Media encompasses 14 newspapers and magazines. We are constantly working to increase the self-service rate on our customer portal, catering to our users’ needs.

Although around 100,000 of our customers use our customer portal every month, we see that more and more people prefer to reach out to us through our chat service. Also, our customers demand longer opening hours and faster response times now than ever before.
We believe chatbots can help our customers in a channel they prefer, providing an immediate customer service presence. Chatbots can assist in solving simple, quick response questions, leaving more time for customer service representatives to focus on complex customer demands and high-touch interactions.
In its first phase, the Schibsted customer care assistant will be able to answer frequently asked questions and guide the user to the right answer. The main focus area will be aiding users with questions regarding subscriptions and logging them into our platform.
Although personality and ability to understand queries is a crucial element in a chatbot’s effectiveness, we also believe a smooth human handover to the Schibsted customer care centre is an important element. Located in Fagernes, Norway, around 100 highly skilled customer care agents are ready to tackle incoming enquiries.
The chatbot will hand over questions that it is not able to answer, dodging situations where a chatbot might not be able to follow up on a customer’s query. When the customer care centre is closed, people will be able to leave a message and then be contacted by an agent during opening hours.
It will be fully transparent to customers whether they are talking to a chatbot or an agent. This will be a key element in the chatbot’s personality. It will not set out to be something it is not, namely a human.
The Schibsted customer care assistant is built in collaboration with Bakken & Bæck on its Orbit Artificial Intelligence (AI) platform. The assistant’s framework is built in Python, where the chatbot’s brain is represented as a finite-state machine.
Transitions between states are based either on natural language processing and deep learning basic analysis of user input, or on hard-coded rules using the expression language Lua. Bakken & Bæck has worked on many AI projects and is a trusted partner when it comes to chatbot development.
Although fully integrated into our stack, the assistant is built on a framework for scalability. Our hope is that learnings from this first phase can be helpful for others in the company seeking the advantages of AI and chatbot technology. Estimated launch of the product is December 2017, so it should be up and running soon.