Custom AI Agents (Coming soon)

Tabidoo AI is designed to be flexible and customizable.

With Custom AI Agents, you will be able to create your own AI assistants tailored to your specific needs.

What is an AI Agent?

An AI Agent is what actually processes your request when you use AI in Tabidoo.

Each agent defines:

• Which LLM model LLM (Large Language Model) is used
• What tools are available
• How the AI should behave (context/instructions)

This allows you to create specialized agents for different use cases.

What will you be able to do?

With custom agents, you will be able to:

• Create agents with specific behavior and knowledge
• Define how AI should respond in your environment
• Connect AI to external systems
• Extend AI capabilities using custom tools

Example use cases:

• Company-specific assistant with internal rules
• AI helper for scripting and development
• Agent specialized for a specific process or domain

How it works

Custom AI Agents are built on three main components:

LLM Model

Each agent uses a selected LLM (Large Language Model).

• Different models are suitable for different tasks
• Models vary in performance, speed, and cost
• You can assign a specific model or use a default one

Tools

By default, an LLM can only generate text based on its training.

It cannot:

• Access external systems
• Read your content
• Execute logic
• Remember previous interactions

Tools extend these capabilities.

With tools, the agent can:

• Load documentation
• Access application structure
• Execute logic
• Call APIs or workflows

Tabidoo provides system tools, and you will be able to create your own.

Context (Instructions)

Each agent includes a context, which defines:

• How the agent should behave
• What rules it should follow
• How it should respond

This acts as a permanent instruction set for the agent.

Types of tools

You will be able to create different types of tools depending on your needs:

URL content

• Load content from websites (e.g. documentation)
• Automatically processed and indexed
• Used by AI to generate answers

Fixed text

• Provide static instructions or knowledge
• Useful for rules, policies, or definitions
• Not sent with every request (more efficient than context)

Script-based tools

• Execute logic using TypeScript
• Return results back to the agent
• Enable dynamic behavior

Workflow tools

• Run full workflows
• Connect to internal or external systems
• Perform complex operations

Agent configuration

Each agent will include configurable properties such as:

• Name & description – what the agent is for
• LLM model – selected model
• Context – behavior instructions
• System tools – built-in capabilities
• Custom tools – your own extensions
• Visibility in chat – whether users can select it
• API access – ability to call agent externally

Extension-based setup

Custom AI Agents will be managed through a dedicated Tabidoo AI extension.

In this extension, you will be able to:

• Create and manage agents
• Configure LLM models
• Define tools
• Test agent behavior

Why it matters

Custom agents make Tabidoo AI:

• More adaptable to your business
• More powerful than generic AI tools
• Fully integrated with your environment