lightbulbKey Concepts

Understand the building blocks of eesel AI

This page explains the core ideas behind eesel AI. Understanding these concepts will help you get the most out of the platform.

Agents

An agent is your AI teammate. You create an agent, train it on your knowledge, and deploy it to handle specific tasks — like answering support tickets or chatting with website visitors.

You can create multiple agents for different roles:

┌─────────────────────────────────────────────────────┐
│                  Your eesel AI Workspace             │
│                                                      │
│   ┌─────────────┐  ┌─────────────┐  ┌────────────┐ │
│   │  Support     │  │  Sales       │  │  Internal  │ │
│   │  Agent       │  │  Chatbot     │  │  Wiki      │ │
│   │             │  │             │  │            │ │
│   │  Zendesk    │  │  Website    │  │  Slack     │ │
│   │  tickets    │  │  chat       │  │  Q&A       │ │
│   └─────────────┘  └─────────────┘  └────────────┘ │
└─────────────────────────────────────────────────────┘

Each agent has its own:

  • Instructions — How it should behave, what tone to use, what to do in specific situations

  • Knowledge — What it knows (your docs, help center, past tickets)

  • Triggers — When it activates

  • Actions — What it can do (reply, tag, assign, etc.)

  • Channels — Where it's deployed (helpdesk, chat widget, Slack)

Learn more: Creating Agents

Knowledge Sources

Knowledge sources are what your agent learns from. The more relevant knowledge you provide, the better your agent performs.

Types of knowledge sources:

Type
Examples
Best for

Help center

Zendesk, Freshdesk, Intercom articles

Customer-facing answers

Past tickets

Resolved conversations from your helpdesk

Learning your team's voice

Documents

Google Drive, Notion, Confluence, PDFs

Internal policies, processes

Website

Any public website

Product info, pricing, FAQs

Macros

Canned responses from your helpdesk

Consistent, pre-approved answers

Knowledge stays up to date — eesel AI automatically re-syncs sources to pick up changes.

Learn more: Training Best Practices

Triggers

A trigger tells your agent when to activate. Without a trigger, your agent waits in the test chat — it won't respond to real tickets or messages on its own.

Common triggers:

Trigger
What it does
Example

Any customer message

Activates on every incoming ticket or message

Full automation

First customer message

Activates only on the first message in a new conversation

Triage and routing

Agent requests help

Activates when a human agent @mentions eesel in an internal note

On-demand AI assist

Scheduled

Runs on a recurring schedule

Periodic reports or checks

You can have multiple triggers on the same agent. For example, you might want your agent to automatically handle new tickets and respond when your team asks for help.

Learn more: Triggers

Actions

Actions are what your agent can do — beyond just generating an answer. Think of them as the tools your agent has at its disposal.

Common actions:

Action
What it does

Leave public reply

Send a response the customer sees

Leave draft reply

Write a draft for your team to review

Leave internal note

Add a note only your team sees

Tag ticket

Add tags for categorization

Assign ticket

Route to a specific team member or group

Close ticket

Mark a ticket as resolved

Create ticket

Open a new ticket

Search tickets

Look up related conversations

Each action can have Human in the Loop turned on, so you can require approval before it executes.

Learn more: Actions

Human in the Loop

Human in the Loop (HITL) is an approval step you can add to any action. When enabled, your agent prepares its response or action but waits for a human to approve it before executing.

This is especially useful when you're first getting started — let the agent draft replies and review them before going fully autonomous.

Learn more: Human in the Loop

Channels

Channels are where your agent is deployed — how customers or team members interact with it.

Channel
Description
Example use

Helpdesk

Agent works inside your helpdesk tool

Zendesk, Freshdesk, Intercom tickets

Chat Bubble

Floating chat button on your website

Customer self-service

Inline Chat

Chat embedded directly in a page

Help center, docs site

Public Chat

Shareable chat link

Email signatures, social media

Slack

Agent in your Slack workspace

Internal team Q&A

You can deploy the same agent to multiple channels — it uses the same knowledge and instructions everywhere.

Learn more: eesel Channels

Instructions

Instructions are how you tell your agent what to do. Think of them as the onboarding document you'd give a new team member — they define your agent's personality, rules, and behavior.

Good instructions include:

  • Who the agent is — "You are a friendly support agent for Acme Corp"

  • How to respond — "Keep answers concise. Use bullet points for steps."

  • What to do in specific situations — "If the customer asks about billing, check their order status first"

  • When to escalate — "If you can't answer confidently, assign to the Support team"

Learn more: Writing Instructions

Integrations

Integrations connect eesel AI to the tools you already use. There are four types:

Type
What they do
Examples

Helpdesks

Where your agent handles tickets

Zendesk, Freshdesk, Intercom, Gorgias, Jira

Knowledge

What your agent learns from

Google Drive, Notion, Confluence, websites

Communication

Where your agent sends messages

Slack, Gmail

E-commerce

Product and order data your agent can look up

Shopify

Learn how to connect your tools: Connect Your Tools

Last updated