Skip to main content

Mode 3: LLM Agent for User Group

Definition​

To complete tasks collaboratively, users with varying abilities frequently need to engage in discussions or coordinate effectively. In this process, LLMs can be leveraged to plan strategies that facilitate both discussion and task delegation within a group. The LLM can help guide the flow of discussion and suggest task assignments among members. The content generated by the LLM agent is then evaluated collectively by the group to ensure alignment and relevance to the task at hand.

LLM Agent in Group Collaboration
Figure: How LLMs enhance group planning and communication.

Key Interaction Modes and Examples​

LLM Agent As An Equal Team member in Single human and Single LLM team.​

Within conversations, LLMs can infer their partner’s intents, needs, and skills by analyzing their actions and communication. They can then adjust their outputs to better align with the common goals of the collaboration. They are capable of correcting errors during task execution refine their actions and communication strategies based on feedback. Instead of directly giving instructions to a LLM, human and LLM are equally capable to complete a task through natural language communication.

LLM agent facilitates the group discussion.​

In this approach, LLMs are used to streamline and facilitate the team’s interaction process, particularly during the facilitating phase. This is typically achieved by integrating an agent within the team that aids in communication, decision-making, information sharing, and coordination. By smoothing out the interaction process, the LLM ensures that the team’s workflow continues seamlessly and effectively, enhancing collaboration and productivity.

LLM agent delegates tasks by analysing members' abilities.​

This approach involves augmenting LLMs with the ability to recognize the unique capabilities of different team members. The primary goal is to leverage the diverse skills of team members to ensure overall team performance, ensuring that tasks are assigned in a way that maximizes each member’s contribution and efficiency. For instance, in response to questions like ‘Can AI help me write better?’, the LLM can decide whether to assign tasks to each member of the group.