The ability of large language models (LLM) to process vast amounts of information, identify patterns, and generate coherent and situated text represents a unique capability that could be advantageously applied in a military context. The use of LLM in military applications presents both opportunities and risks, which necessitates a thorough understanding of their impact on the sociotechnical systems in which they are deployed. This report investigates how LLM can be used in military tactical planning and what effects their introduction has on the planning process, as well as users' attitudes toward their use. Two studies were conducted using a custom-instructed language model (hereafter referred to as the AI tool), adapted for a military setting and instructed in the "Planning Under Time Pressure" (PUT) methodology. The first study was conducted at the non-commissioned officers training program at the Military Academy in Halmstad with ten officers, and the second during the Tactical Army Course at the Land Warfare Centre in Skövde with 61 officers. Data collection included focus group interviews, surveys, and observations. Twelve aspects in the form of experiences, lessons learned, or perspectives on the use of large language models in military tactical planning were identified. Expectations for AI systems were high, although participants felt that the AI tool contributed meaningfully only to a limited number of tasks. It was particularly useful for summarizing large volumes of text and to some extent for generating courses of action. However, the tool performed poorly when supporting maprelated tasks, which limits its usefulness for spatial analysis. The studies demonstrate that LLM can be applied to tactical planning, albeit with some limitations due to current technical constraints. A prerequisite for AI tools to streamline tactical planning is that they are integrated into a digitalized command-and-control system with access to data streams from sensors on the battlefield. Further research is needed to better understand the effects of using LLM in tactical planning and to determine how AI tools should be designed to provide the most effective support.