Emely is a virtual dialogue robot that is available around the clock for people who want to practice job interviews but also practice their language skills in Swedish.


Being interviewed for a new job is a tough challenge for many. For some it may be even their first job interview, for someone it is difficult to give detailed and good answers when Swedish is not my first language, and for someone third it may be both of these at the same time.

Imagine if you could practice understanding questions and giving thoughtful answers. If that by practicing job interviews can make it possible to give a more accurate impression of oneself. It will also be easier for the employer to find the right skills for a job, instead of just taking the one who happened to be the best at being interviewed.

In the next step in the development, we will, in collaboration with the Health and Care Administration, train Emely for conversations in the care professions. Emely becomes a tool where the user trains his Swedish before a job and during a job.



Emely will simulate a job interview with the user, by asking questions and follow-up questions that are relevant to the job the user has applied for. The user is expected to respond to them in speech, to practice the sharp situation that a job interview is. The goal of Emely is to make the user feel comfortable in the situation and that he or she answers questions related to the service they could apply for.

Emely's language intelligence is based on machine learning algorithms (ML algorithms), which belong to artificial intelligence (AI). This sets Emely apart from most of the classic chatbots and chatbots that are on the market today, which are rule-based and that can not handle major extravagances from their pre-programmed topic. ML algorithms are mathematical functions that are adapted to historical data for different tasks, for example to be able to predict (guess at / estimate) future outcomes of a question or to create a basis for decisions. In the field, the actual adaptation of the algorithm to historical data is called training.



Impact target

More people will get to work with the help of Emely.

Resident involvement

Emely develops iteratively in close collaboration with the end user.


Name: Peter Ekström
E-mail: peter.ekstrom@helsingborg.se