AI improves wellbeing and safety at work
Artificial intelligence is already being used in many workplaces to search for and organise information. It has enabled us to develop wellbeing at work at Posti and Foibekartano and improve occupational safety at Outokumpu’s factory in Tornio.
AI is changing the way companies operate, make decisions and serve their customers. It can automate tasks, provide insights and increase productivity. Artificial intelligence can perform work that is too dangerous, heavy or time-consuming for humans.
In expert work, many have already tested artificial intelligence in searching for information; answers are available quickly and clearly as long as you know how to ask the right questions. If your ideas are not moving forward, artificial intelligence can help, even if it does not provide ready answers.
You can discuss your tasks and perspectives with AI.
“For example, Microsoft’s Copilot, which is available to Posti’s experts, provides tips to help them move forward in their work, but humans are responsible for the actual problem solving,” says Heini Toivonen, Vice President, HR Development and Operations at Posti.
Artificial intelligence does not remove the importance of the customer advisor, but it enables people to solve problems faster and more efficiently.
A pilot project has been launched in Posti’s customer service to reduce the workload of employees by utilising artificial intelligence to search for and process large amounts of information quickly. Artificial intelligence does not remove the importance of the customer advisor, but it enables people to solve problems faster and more efficiently.
“A customer advisor meets hundreds of people every day and needs a lot of information in order to be able to serve them. Searching for information increases the employee’s workload, while they should be able to focus on genuinely encountering the customer via chat or on the phone,” Toivonen describes.
In the pilot for Posti, the AI solution is called support intelligence because it helps the customer advisors find the right instructions related to Posti’s services and formulate answers in clear language based on them.
“The customer advisor must be able to ask the right questions from AI in order to get the right answers. We train both the machine and its users at the same time,” says Toivonen.
AI can even interpret tone of voice
Artificial intelligence is also used in many other tasks at Posti.
The Peakon system, which supports the collection and listening to employee feedback, is used twice a year to conduct a personnel survey that provides supervisors with numerical information on their employees’ wellbeing at work, job satisfaction, commitment and supervisor experience.
Peakon gives suggestions for things to pay attention to, i.e. topics that many employees mention in their answers. The manager will receive the results of the surveys if there are at least five respondents, and anonymity is guaranteed.
“I see the answers of my 30 employees in Peakon, analysed by artificial intelligence, but I cannot identify individual respondents,” Toivonen says.
We have received even more precise data on what we need to focus on in order to improve wellbeing at work and employee satisfaction.
Posti has also utilised an AI solution that can interpret responses in greater depth and even identify the employees’ “tone of voice”.
“We have received even more precise data on what we need to focus on in order to improve wellbeing at work and employee satisfaction,” Toivonen says.
Posti has particularly high expectations for the benefits of artificial intelligence in recruitment during peak seasons.
In the near future, artificial intelligence can streamline job seeking and improve the job seeker experience.
Every year before the holiday season, Posti receives a five-digit number of seasonal assistant applications. In order to facilitate the peak recruitment season, Posti has studied an AI solution where a virtual recruiter could chat with the applicants, for example, by text message or messaging apps, and provide additional information about the nature of the work, location and working hours.
The virtual recruiter is proficient in many languages and can book appointments in the applicants’ calendars.
“In the near future, artificial intelligence can streamline job seeking and improve the job seeker experience, while at the same time freeing up recruitment professionals to interview and assess the applicants,” Toivonen says.
AI robot collects occupational safety data
Outokumpu’s ferrochrome mill in Tornio uses artificial intelligence for occupational health and safety monitoring.
The intelligent and learning Jokkeri robot is tasked with preventing noise, dust and vibration to improve the working conditions of employees. It smells gases, detects sounds and measures temperature.
“We have gained a huge amount of new information, which is stored in the database and which we can use to develop occupational safety, for example. The development of the robot and teaching of artificial intelligence are still in progress, but the results are already very promising,” says Juhani Keinänen, Occupational Safety and Health Ombudsman.
A 3D camera scans the work area from floor to ceiling and wall to wall as the robot moves around the site.
Each task is taught to Jokkeri individually and, once it has learned it, it can measure sound waves, identify gases or climb steep stairs, for example. A 3D camera scans the work area from floor to ceiling and wall to wall as the robot moves around the site.
“It can learn from its own mistakes, just like humans. Once, there was frozen water on the floor and Jokkeri hobbled over the slippery spot. The next time, it had already learned how to move on slippery surfaces,” Keinänen enthuses.
He hopes that at some point, the robot could move around the workspaces completely independently, but for the time being, it requires monitoring by humans.
“Our working environment is challenging for both machines and people. If a robot can cope here in the product processing plant, it can cope anywhere,” says Keinänen.
However, Jokkeri differs from its predecessor in being a real artificial intelligence robot.
Juhani Keinänen has always been interested in robots and he flies drones in his spare time. Nevertheless, the 58-year-old employee had his doubts about working with an AI robot.
“Well, I will not become a coder anymore, but I am doing well with Jokkeri. I received two weeks of instruction before the implementation. Our electricians code the robot, but it waits for its tasks in my room.
Robots are not new at Outokumpu’s factory, as they have been used there since the 1990s. However, Jokkeri differs from its predecessor in being a real artificial intelligence robot. And best of all, its data is never lost, even if the machine itself breaks down.
“Everything that Jokkeri learns can be transferred to the new device, i.e. the data is never lost. In the future, similar robots will probably utilise swarm intelligence, learning everything from each other,” Keinänen predicts.
AI training on the job
Employers hope artificial intelligence will become a tool that makes work more efficient, but if used poorly, it can have the opposite effect.
In a study coordinated by the Kalevi Sorsa Foundation, artificial intelligence planned employees’ work shifts to within an accuracy of 15 minutes based on cash register data in a retail unit. However, the algorithms did not take into account invisible work, such as filling magazine shelves, and it called to work too few employees. This resulted in chaos and customer dissatisfaction.
However, when used correctly, artificial intelligence is beneficial. According to the study, it is essential that employees are involved in the planning of new information systems from the start.
New technology should always be implemented gradually so that we can train ourselves and learn every day together.
This is precisely what has been done at Foibekartano, a private housing service company for the elderly in Asola, Vantaa. Its operating principle is to develop the digital skills of both employees and residents together, through trial and error.
“New technology should always be implemented gradually so that we can train ourselves and learn every day together. Training outside the company should also be carried out in groups, so that development can start immediately,” says CEO Ulla Broms.
Foibekartano has participated in Haaga-Helia’s AI-TIE artificial intelligence programme, one of the aims of which is to facilitate time-consuming work time planning. The experiment, which began two years ago, has already freed up a lot of employees’ hours for the most important thing: meeting customers.
The system has been taught to understand that Foibekartano’s employees have round-the-clock working hours and that there must be a certain number of employees and the right expertise present to meet the personnel requirements, such as a sufficient number of medicine rights holders.
Artificial intelligence has also been taught to take into account the employees’ wishes for days off, working hours restrictions and areas of expertise. It has enabled the company to cut unnecessary overtime and reduce the need for substitutes.
The impartiality of artificial intelligence increases trust
Broms considers it very important that the personnel gradually adopt the new way of work shift planning and participate in its development. Some residents have also been interested in following the process.
“When elderly people and staff are not afraid of new technology and try it out, even equipment manufacturers are willing to ask us for information, tips and development suggestions.
According to Broms, one significant advantage of AI-based working time planning is impartiality: people can consciously or unconsciously try to control other people, but artificial intelligence is impartial.
“Impartiality increases mutual trust in the long term and also the well-being of the personnel,” says Broms.
AI can also be of service in situations where a nurse does not speak Finnish as their native language.
Foibekartano believes that artificial intelligence can support many work tasks in the future. Good experience has been gained in the field of recording customer and patient data using speech recognition. AI can also be of service in situations where a nurse does not speak Finnish as their native language.
“Artificial intelligence could also help in the analysis of health data so that the diet of the elderly can be planned in an ideal way and the treatment of chronic diseases can be more individual,” Broms plans.
Anna palautetta jutusta