|
minute read
Download Client Story
Background
BOSA, the Federal Public Service for Policy and Support, provides supporting services to the Belgian government and the federal institutions in various areas: IT, HR & recruitment, organisational management and integrity policy. Furthermore, it has
an oversight role for budget and accounting. It operates under the supervision of the Secretary of State for the Digital Agenda, Digitization and Privacy who is linked to the Prime Minister.
Within BOSA, the Directorate-General Digital Transformation (DGDT) leads and promotes, amongst other things, the digital transformation of federal institutions by providing advice and developing projects in connection with new technologies, adopting a
citizen-centric approach whenever possible.
In this context, BOSA began experimenting with chatbot technology. The first selected use case consisted in assisting candidates applying for a job in the federal public sector through the Selor recruitment website. The chatbot will at some point be integrated
into the Selor website and work in complement to the existing Q&A and the human-operated service desk which can be reached via phone, mail or a web form. Selor handles more than 20.000 requests on a yearly basis and that number grows steadily.
The Challenge
BOSA wanted to experiment rapidly with the latest chatbot technologies and understand the steps required to build and maintain a virtual assistant. This meant that business owners, namely the first-line helpdesk, the HR experts and the communication department,
should be able to take ownership of the solution in order to ensure its long-term success.
The Solution
Sopra Steria leveraged its own solution Alive Intelligence, which provides an intuitive user-friendly cloud-based platform allowing business users to configure, train, test and maintain the conversation scenarios that are used by the chatbot.
Alive Intelligence provides an abstraction layer on top of the AI engines such as IBM Watson or Rasa, avoiding vendor locking and allowing BOSA to retain control and ownership on the implemented conversations.
Sopra Steria proposed a global agile approach, including:
- co-design workshops to refine the use case and define the backlog of conversation scenarios and key configuration variables of the chatbot,
- training sessions to give the ability to BOSA staff to configure and maintain the conversations up to date
- the implementation of the Minimum Viable Product (MVP) by its own consultants
- technical and project management support during the sprints to allow BOSA to build the pilot
- support to Alive Intelligence deployment and operations
