Harnessing AI: LLM Pilot Enables Smarter, Transparent Decision Support

The Data for Efficiency, Productivity and Competitiveness project has launched a pilot with SK Protect Oy, which aims to explore ways to utilize large language models (LLM) in intelligent, data-driven applications. Protect is a forerunner and a leading partner in services that relate to safety and quality.

The pilot tests ways to utilize large language models as part of the software architecture, for example, in collecting and processing observations and deviations, in processing information, and as decision-making support in intelligent systems.

The goal is also to find solution models in which the language model agent does not operate completely independently, but the user can control and check its operation. This enables the controlled and transparent utilization of artificial intelligence in, for example, the following situations:

• Data collection and classification: The application can chat with the user and collect information that the language model automatically classifies, for example, by subject area, customer type, or urgency.
• Investigation: LLM can assist the user in determining the root cause and suggest corrective actions.
• Summary generation: LLM can analyze the data collected in the application and produce summarized information to support decision-making.
• Result storage and further use: The results produced by the model can be stored in a database, from which they can be further utilized, for example, in reporting or analytics, or as input to a language model to produce new information.

This approach can combine the efficiency of artificial intelligence and human expertise, enabling intelligent but manageable solutions for data utilization.