Artificial intelligence is now also at work in Vibemetrics
Artificial intelligence and machine learning is one of the biggest megatrends of our time. The use of AI offers a way to analyse huge amounts of data efficiently. AI is also used in the Vibemetrics service. We interviewed Vibemetrics' summer intern Olli Ruokojoke, who develops analytics and the use of AI in employee surveys.
Who are you and what do you do?
I am Olli Ruokojoki, a 21-year-old student of engineering. I am a second-year student of engineering physics and mathematics at Aalto University School of Basic Sciences. I've been working at Vibemetrics for three summers, where I develop comment analysis and the artificial intelligence behind comments.
Agile learning has been much talked about recently. Agile learning differs from traditional predictive learning in that it seeks new knowledge in a rapid cycle to support the work being done. Learning is sought through a wide range of channels, and qualifications and formal education in general play a less important role. Our summer intern Olli is a true master of agile learning!
I've practiced coding quite a bit on my own by studying and doing. I became interested in AI and analytics a few years ago, especially through Max Tegmark's book "Life 3.0: Being Human in the Age of Artificial Intelligence". I've been building up my knowledge through a university coding course, watching AI-related videos and listening to audio books on the subject. In high school, I wrote several essays on AI, including an English exam. I am planning to minor in computer science for my university studies.
What has the development of AI meant in practice?
I am developing the analysis of comments and the AI behind the comments. I have been studying different algorithms to analyse comments and have chosen the most suitable ones for us and Miko to use. I have created a program that finds similar keywords in the data. For example, the program can match "nice colleague" with "nice colleague", meaning the same thing. When there are a large number of keywords, the program can combine words that mean the same thing under the same main keyword, making it easier to go through them. In this way, important entities can be found in the data.
How can AI improve the service and help the users of ViBe?
AI will improve and streamline the use of comments and open responses. Employee surveys can have thousands of respondents and several thousand open comments. It is challenging if, for example, one person has to go through all the comments. It is difficult to get an overview of the large amount of data and to separate the important issues from the mass. AI helps to analyse the results by spotting certain recurring themes and keywords that make it easy to see where things stand and where to focus. This helps to identify problem areas and what is good and working based on the comments, as well as what comments are worth responding to. AI is used to find the relevant issues in the sea of comments.
"Developing AI requires deep understanding and creativity"
Mikko, you were pretty excited when you came out of the meeting with Olli. What got you so excited?
Well yes! I've learned a lot about AI with Olli. It's funny to see that it's not really about some mystical thing, but simply about telling the machine what to do when various conditions are met. So, in effect, we are telling it what to do when it detects certain words or certain phrases. At this stage of its intelligence, the machine cannot yet create condition clauses by itself, but must be taught to do so. Formulating good conditional clauses requires an understanding of what is important in human life, management and organisational development. It requires good skills and is ultimately quite a creative job. I'm really quite excited about it!
The use of AI requires a large number of comments. With a small number, the machine gets confused and can't tell the important from the trivial. We're still working on it and learning, but the AI feature is already operational in the service.