For performing in project management, PTA is investigating the possible applications of AI in project work. When considering the most likely forms of AI application in everyday project work, PTA distinguishs between tasks of varying complexity, from simple to complex. Therefore, when assessing the possible applications of AI from this perspective, a distinction must be made between the use of language models and AI agents.
Supplement
The aim of the feasibility analysis is to increase the efficiency and quality of our employees' project management. For the analysis, specific areas of application for AI in project management are identified and evaluated in terms of their feasibility and relevance. Use cases are then derived and tested using suitable AI solutions. Language models (LLM) handle simple tasks, while AI agents are used for more demanding tasks.
Subject description
Language models, also known as large language models (LLMs) such as GPT4, are special AI models that have been trained exclusively on text. They have billions of parameters and can generate answers in natural language, summarize texts, or write code. An LLM is therefore a very powerful knowledge generator and language generator, but it is not in itself an independent actor that pursues goals or performs actions. An AI agent is a system that not only uses a model (e.g., an LLM) but also implements behavior. It observes its environment, plans steps, executes actions, and learns from the results in order to achieve a defined goal. An agent can be thought of as a small robot in the software universe that uses an LLM for language processing, but also has logic, memory, and decision-making mechanisms.