The potential of AI is enormous – but quality, ethics and humanity make the difference
2M-IT’s Application Services provides support for the systems and applications used by wellbeing services counties. Business Director Kati Pöllänen sees AI above all as a tool that improves the flow of work and opens up new possibilities.
“AI solutions serve us by making our own work faster and easier. At the same time, AI supports experts in searching for and combining information, helping us find solutions and perspectives that we might not otherwise notice.”
The potential is enormous, but reliable benefits arise only when three things are in place: the data is of high quality, its use is guided ethically and securely, and people remain in control of both decision-making and quality assurance.
AI needs the right kind of high-quality source data. That is why careful and structured documentation is essential: if the source information is incomplete, the suggestions may also lead in the wrong direction. AI-generated suggestions must always be evaluated and approved by a human.
AI in incident management and service management
In the ServiceNow service management system introduced at 2M-IT in 2024, there is an AI assistant called Now Assist. It helps identify and handle incidents, analyzes historical data, suggests solution options, and produces log summaries and final reports. This speeds up response times and helps detect anomalies earlier than before.
In incident management, AI is used, for example, to support the leadership of a situation room – especially in compiling logs and reporting.
“AI helps us quickly identify and prioritize incidents, as well as compile and analyze information from previous cases.”
In service management, AI supports tasks such as reviewing billing and contracts, preparing presentation materials, and compiling meeting notes, summaries and proposed actions. The goal is to leave experts with more time for assessment, decision-making and collaboration.
Data quality: the foundation of trustworthy AI
The reliability of AI-generated solutions depends directly on the data. That is why clear criteria must be set for the structure and content of the data, and day-to-day work must emphasize careful documentation, consistent practices and accuracy: even small gaps or errors can quickly scale through AI.
“The most critical quality assurance step before introducing AI is data quality. We need to know exactly what data the AI is using – and every expert must be precise in collecting and storing that data.”
Ethics and responsible use: people in control
Responsible use of AI requires clear policies: what data may be used, for what purpose, and on what basis. When boundaries, responsibilities and approval practices are transparent and commonly understood, it is easier to trust AI – and risks (such as misuse, bias and data protection risks) remain under control.
The discussion around agents and advanced automation has accelerated: there is talk of agents replacing human work and other agents monitoring their work. However, it is important to remember that AI draws conclusions from existing data. If errors or gaps enter the data, they may be carried into the solutions – and in the worst case, begin to multiply if the outputs are used as the basis for new data or decisions. That is why people also play a key role in quality assurance of AI-generated solutions.
Learning and a culture of experimentation: permission to try – and fail
When the boundary conditions (data quality, policies and human oversight) are in place, the potential is enormous. That is why we should encourage people to learn, experiment and share their experiences of using AI. A safe atmosphere also means that people can learn from failures without fear.
Humans are not disappearing: AI does not replace interaction
AI can speed up information processing and support decision-making, but it does not replace human interaction, trust or collaboration. Complex situations are often resolved through discussion, by combining experience and tacit knowledge, and by making well-considered choices together.
That is why scenarios in which agents would one day handle Teams meetings on behalf of people sound, at the very least, unfamiliar: technology can support meetings, but meaning still comes from people meeting and building shared understanding.
Summary
AI can significantly enhance work efficiency and help find solutions faster. However, the benefits are reliable only if the underlying data is of high quality—which requires careful and consistent documentation in everyday work. In addition, clear ethical ground rules are needed for the use of data, and people must remain in control even as agents and automation increase; otherwise, errors can be carried into solutions and multiplied. When these conditions are in place, experimentation and learning should be encouraged—but it is important to remember that AI supports human collaboration, it does not replace interaction.
You might be interested in learning more: