Communication channels are becoming increasingly digital. Businesses and public authorities receive many enquiries via social media, their websites, email or telephone. To ease the burden on staff when dealing with these enquiries, more and more businesses and public authorities are using chatbots and voicebots. It is not enough simply to create the bot. Rather, it must be continuously trained and optimised. The biggest challenge here is operating the respective systems, particularly when this requires specialist IT knowledge. If a company wishes to deploy multiple assistants across different areas, maintaining and expanding the bots takes up a great deal of time. It is precisely this maintenance that is crucial to ensure the bot can continually cope with new requirements and challenges. To enable even editors without IT knowledge to manage this, Materna has developed a user-friendly chatbot content management system.
A comprehensive solution with many advantages
The chatbot content management system is intuitive and can be used without any IT experience. As it is less complex, there is no need for lengthy training periods. Furthermore, the content management system enables the creation and simultaneous maintenance of one or more digital assistants. Even multiple bots – whether chatbots or voicebots – can be clearly consolidated and tested within a single system. The dual-control principle for approval processes ensures high quality and reduces the error rate. Under this principle, for example, any change can be tested using a preview function before publication and improved if necessary. In general, the system is a comprehensive, customisable end-to-end solution in the chatbot sector.
In use at KfW
The example of KfW illustrates the benefits offered by the editorial system: the development bank offers students a special student loan. As with most financial matters, there is a high demand for information. Enquiries peak particularly at the start of the academic term, placing a heavy burden on KfW staff. To support both its own staff and students, KfW sought a suitable chatbot solution. It found what it was looking for with Materna and IBM Watson Assistant, which is accessed via the IBM Cloud. Materna provides the full service (managed service) and support, including ticket handling. The chatbot is fed via the content management system.
Since October 2019, the chatbot has been providing students with information on key aspects of student loans. For example, the chatbot provides answers to frequently asked questions: Who is actually eligible for a loan? What are the repayment terms? How is the student loan repaid? This is a particularly common question during the coronavirus crisis: as the German government has changed the terms and conditions of student loans to support students, interest in student loans – and consequently demand for the product – has risen enormously.
Easy management of the chatbot via the content management system
An important feature of the chatbot content management system is that editors can not only enter simple responses to user queries, but also have entire dialogue trees at their disposal. In this way, just as in a human conversation, editors can store as many follow-up questions and meaningful answers as they wish within the system. Existing information is linked quickly and easily within the dialogue flow.
For KfW, this means it can quickly create new topics and combine them with existing content. Furthermore, editors can retrain the bot independently and directly within the system. To enable the chatbot to be further improved, the editorial front end includes a reporting section with a dashboard. The reporting section helps, for example, to present elements to be evaluated in graphical and tabular form. Here, it is possible to analyse the recognition rate in the dialogues, determine whether dialogues need optimising, and identify where further question variants or synonyms are required.
Support from Materna
Materna supports companies and public authorities in implementing the chatbot editorial system and, where required, customises the system, adds additional features, assists with editorial services, advises on other matters and integrates external systems. The system, which is specifically tailored to the needs of content editors, enables the management of all chatbot metadata and the maintenance of the bot in one place. In doing so, the solution removes many of the hurdles that companies would otherwise face when setting up and maintaining chatbots.
