1. Information about the procurement
ICT Solution: “Chatbot Albert”
The Austrian Patent Office aims to further develop the customer service “Chatbot Albert” to become more intelligent and more user-friendly. Currently, “Albert” helps customers and interested citizens via the website www.patentamt.at and via Facebook. He answers questions regarding brands, patents, fees, funding and seminars. However, “Albert” cannot process natural language and cannot form full sentences on his own. He reacts mainly to keywords and answers questions by posting pre-formulated answers out of a “MySQL” database or the SOLR-search of the website www.patentamt.at. An additional problem is that adding new information or content to the database is a complex process for the editors of the Austrian Patent Office.
The aspired results are:
- Backend-solutions that enable “Albert” to interpret natural language and form answers. It should be possible to integrate new features from time to time – “Albert” wants to develop. An information transfer between “befriended” chatbots would be ideal (e.g. with the chatbot of the City of Vienna).
- A very important factor is the editing of information by the Austrian Patent Office’s staff via an innovative and user-friendly CMS system! Editing, maintenance, regulating and testing should be possible within this system. It would be possible to extend the existing system or replace it completely.
- Analysis of behavioural user-data should be possible.
- A so-called human handover must be possible for emergency-situations or technical difficulties. In one of those cases, Albert should alert one of the editors automatically.
- Extension or replacement of backend and frontend of the chatbot-system.
The innovation procurement of the Austrian Patent Office is planned under the national digital/ICT strategy. This digitalization project will be reported to the Ministry for Transport, Innovation and Technology.
- Artificial Intelligence, machine learning, understand (NLU) and process (NLP) natural language.
- Content management system (CMS).
- Behavioural user-data analysis.
- Hosting of the system.