Agents Overview

---

What are agents?

Agents are the AI tools you build within the Pickaxe platform. Each agent is a standalone AI tool designed to serve a specific purpose — defined by the instructions you give it.

Note: You may see agents referred to as "pickaxes" in some places — that's what we used to call them. Same concept, new name.

You can think of an agent as a custom AI assistant that knows exactly what you need it to do. Whether it's answering customer questions, generating content, qualifying leads, or walking users through a process, every agent is shaped by how you configure it.

The Agents dashboard

The Agents dashboard is your home base for managing all of your agents. From here you can see every agent in your workspace at a glance, along with key details like the model it's using, how many deployments it has, its total usage, and when it was last updated.

At the top of the dashboard, summary cards show your workspace-level stats for a selected time period — including total cost, number of users, and total runs — so you can quickly monitor activity across all of your agents.

You can search, filter, and sort your agents to find what you need, and use bulk actions by selecting multiple agents at once.

What makes up an agent?

Every agent is built from a few core components that work together:

Instructions

Instructions are the foundation of your agent. They tell the agent what its purpose is, how it should behave, what tone to use, and any rules or boundaries it should follow. The more specific your instructions, the more focused and useful your agent will be.

Knowledge base

You can train an agent on your own data by adding content to its knowledge base. Upload documents, paste text, or point to URLs — your agent will use this information to give accurate, grounded responses based on what you've provided.

Actions

Actions let your agent do more than just chat. You can connect it to external tools and APIs, or even to other agents, so it can take real actions like looking up data, submitting forms, sending emails, or triggering workflows.

Model

Each agent is powered by an AI model. You can choose from a range of leading models and switch between them at any time to find the best balance of quality, speed, and cost for your use case.

Deployments

Once your agent is ready, you can deploy it in multiple ways — embed it on a website, share it via a direct link, add it to a portal, connect it as a Slack bot, WhatsApp bot, email bot, or access it through an API. A single agent can be deployed across as many channels as you need.

Creating a new agent

To create a new agent, click the + New Agent button in the top right corner of the dashboard. This will open the Agent Builder, where you can configure your agent's instructions, knowledge, actions, model, and more.

You can also browse the Template Library to start from a pre-built template created by the community.

Managing your agents

From the dashboard, you can:

• Search for a specific agent by name
• Filter and sort to organize your list by model, usage, or last updated date
• Select multiple agents to perform bulk actions
• Click on any agent to open it in the builder and edit its configuration
• Access quick actions using the menu icon on the right side of each row