Schibsted

Schibsted_Media_Group_2010_1
Client
Schibsted
Delivery
Design, frontend and machine learning
Period
2017 - 2018

Every year Schibsted gets 1,3 million inquiries from customers, divided among their support team of 113 customer service representatives. They came to us with an ambition to ease the load of the most repetetive and frequent requests, which would make their staff able to provide more quality and value to their support.

The solution was a scalable chatbot framework that handles requests 24/7.

The chatbot answers FAQs and provides assistance by performing simple customer service tasks, like solving login issues, sharing digital subscriptions and changing the delivery address for the newspapers. The bot is integrated with Schibsted’s call center technology stack.

The bot is built on top of our Orbit Operator platform, a framework for building custom virtual assistants that can be integrated into enterprise systems.

The solution consists of

  • A chat widget. A custom chat interface that is integrated with the different newspapers’ account pages. The widget is tweaked to fit the different brand styles and data privacy policies.
  • Trained machine-learning models for understanding user intents. The bot was trained on data from more than 80 000 chats to identify frequent questions and answers. It continues to learn every day.
  • An admin panel that lets Schibsted train the chatbot and set up dialogue templates, in order to effectively pass on their expertise (and boring tasks) to a digital assistant.
  • Conversation flow logic that is customised with customer specific elements like predefined responses, steps for process automation and business rules.
  • Human handover. We all know how frustrating customer service conversations can be when the bot doesn’t get it right the first, second or eleventh time. So we made it easy for users to speak to a human being in case they need to.
  • Hooks for integration with customer care technology stacks, in Schibsted’s case Siebel CRM and the call center solution Puzzel.
  • Customer insight gathered from Chatbase analytics.

In the beginning of its career the chatbot would only be live during work hours because of the need to have people at work ready for an eventual handover. Lately the bot was put live 24/7, a true testament to its efficiency and value.