Think of an agent as a specialized team member with specific skills and expertise. AI Agents are the building blocks of your future AI workforce. The AI workforce that you will build and deploy within Lampi AI.
Here's are some considerations when working with AI agents.
Accessing and launching agents
You can access your Agents directly from any chat by clicking on the "Agents" modal. Once in this section, you will see a list of available agents to choose from. To initiate an agent, simply select the one you wish to use and click the launch button.
Once launched, the agent operates in the background, allowing you to continue your work without interruption. This means you can focus on other tasks while the agent processes the workflow in parallel. This feature is particularly beneficial for managing multiple tasks at once, as it helps you to maximize your productivity.
You can launch multiple agents simultaneously.
Queuing system for agents
Lampi AI employs a queuing system for the agents. This means that if you launch several agents at the same time, they will be executed in the order they were initiated. The system ensures that each agent receives the necessary resources and processing time to complete its assigned task.
Monitoring agent status
To check the status of your agents, navigate to the "Tasks" tab in the menu. Here, you will find a comprehensive overview of all your launched agents, including their current state—whether they are running, completed, or queued. This feature allows you to monitor the progress of your tasks easily.
Managing your agents
In the "Tasks" tab, you also have the ability to manage your agents effectively. For instance, if you realize that an agent was launched with incorrect files or parameters, you can delete it from the task list. This flexibility ensures that you maintain control over your workflows and can rectify any mistakes promptly.
Duration of an agent
The duration of a task completion is primarily influenced by the number of prompts and the tools utilized to execute them.
For instance, if you decide to ask the agent ten prompts with the tool "Web Search", it will take the time to process each web search, including processing data, navigating through various web pages, and gathering relevant information. Additionally, there are further steps associated with the agent's mode of operation, such as iteration, result analysis, and information condensation, which can extend the overall process. Each of these stages requires a specific amount of time to execute.