Subtitle: dynamic on-the-fly access to (new) data sources and web services by chatbots

The Web is facing a major paradigm shift. Over the next few years it will transform from a human readable platform to a platform for bots. Therefore, the prerequisite for content, data, and service providers to be visible in this "new" web, is the proper structuring and the semantic annotation of their data and services. Semantic technologies will be the backbone infrastructure enabling this paradigm shift.

Chatbots and intelligent personal assistants currently only work with hard wired data sources and services. They can give information about addresses, opening hours, the weather, a recipe and more. They can also use services, like maybe a table reservation or the purchase of a movie ticket. But only under the premise that the data source or services is known and the connection is implemented by the developers (hard wired). What they can not do up to now is accessing Web data sources on the fly or going into interaction with Web services dynamically. If an assistant wants to use data from a web source, it has to understand the structure in which the data is present. And if an assistant wants to use a Web Service, it has to know its address, the input parameters, the expected output and more. If this information is not available in a machine readable way (semantically annotated), the assistant can not interact.

Currently, the main problem is that neither data on the web, nor services on the web offer a machine readable and machine understandable semantics. So to be able to interact with web data sources and web services on the fly, the assistants would need to understand the structure in which the data is present and the protocol in which a service can be accessed or consumed.

Each chatbot or voice assistant needs a lot of service integrations to fulfill requests and tasks. Therefore a scalable methodology for service integration and execution is needed. Semantic technologies will help assistants to perform service calls. The goals of SDB2 are therefore: 1) to annotate touristic service providers' APIs: to make Web Services machine read- and interpretable, they have to be described in a common language or with a common vocabulary. One goal of that project is therefore to annotate touristic service provider's APIs with the vocabulary of schema.org. We will analyze a selection of APIs and develop a best practice for API annotation. And 2) integrate these services into a platform for an easy integration in chatbots and voice assistants. It is expected that the developed methodology of this project leads to a faster scalability of service integrations in chatbots and voice assistants including an automation of the underlying dialogues for the service processing. Existing chat bots will be extended with the ability to read, interpret, understand and execute annotated web services.