Session Todo Lists

When AI agents tackle complex requests, they often need to break work into steps, track progress, and maintain visibility into what remains to be done. ChatBotKit's Session Todo Lists provide exactly this capability - a structured way for AI agents to create, manage, and track todo items during conversations, enabling more organized and transparent multi-step work.

Without task tracking, AI agents working on complex problems can lose track of progress, skip steps, or struggle to communicate what has been completed versus what remains. Session Todo Lists solve this by giving agents a persistent (within the session) mechanism for structured task management that both the agent and user can reference throughout the conversation.

Key Capabilities

Structured Task Management

Your AI agents can create todo lists with individual items, each containing a unique identifier, a concise title, and a status. The three status options - not-started, in-progress, and completed - provide clear visibility into task progress. This structure helps agents organize complex work into manageable, trackable units.

Session-Scoped Persistence

Todo lists persist throughout the conversation session, stored securely and automatically cleaned up after 24 hours of inactivity. This means your AI can reference and update tasks across multiple exchanges within a conversation without losing context, while avoiding permanent storage of temporary planning artifacts.

Complete List Replacement

When updating todos, agents provide the complete current state of the list rather than individual modifications. This approach ensures consistency and makes it easy to add new items, update statuses, or remove completed tasks in a single operation. The agent always works with the full picture of what needs to be done.

Namespace Support

Todo lists can be scoped to different namespaces, allowing agents to maintain separate task lists for different contexts within the same session. This is useful when an AI is helping with multiple distinct workstreams or when different parts of a workflow need independent task tracking.

Real-World Use Cases

Multi-Step Research Tasks

When a user asks an AI agent to research a topic comprehensively, the agent can create a todo list breaking down the research into steps: gather sources, analyze each source, synthesize findings, draft summary. As the agent works through each step, it updates the status, providing the user visibility into research progress and what remains.

Document Processing Workflows

An AI helping process multiple documents can track each document as a todo item. Starting with all items as not-started, the agent marks each as in-progress while working on it and completed when finished. If interrupted or if the user asks about status, the todo list provides an instant snapshot of progress.

Complex Problem Solving

For technical troubleshooting or debugging tasks, an AI can maintain a checklist of diagnostic steps to attempt. This ensures systematic coverage of possibilities and lets the user see exactly what has been tried, what is currently being investigated, and what diagnostic options remain.

Project Planning Assistance

When helping users plan projects, AI agents can create structured todo lists representing project phases or deliverables. The agent can help refine the list, reorder priorities, and track which planning elements have been addressed during the conversation.

How It Works

Session Todo Lists are available through the todo management abilities in your skillset configuration:

• Read Operation: Retrieves the current todo list, returning all items with their IDs, titles, and statuses. Returns an empty list if no todos exist yet.

• Write Operation: Replaces the entire todo list with a new state. Your agent provides the complete list including any existing items that should be preserved along with updates or new items.

• Manage Operation: A unified operation that accepts either read or write as a parameter, allowing flexible integration into agent workflows.

Todo items follow a simple structure:

• ID: A numeric identifier for tracking individual items
• Title: A concise, action-oriented description (ideally 3-7 words)
• Status: One of not-started, in-progress, or completed

The todo list is automatically stored using the conversation session context and expires after 24 hours of inactivity, ensuring temporary planning data does not persist indefinitely.

Getting Started

To enable Session Todo Lists for your AI agents:

1. Navigate to your skillset configuration in the ChatBotKit dashboard
2. Add the todo management abilities to your skillset
3. Update your AI's instructions to describe when and how it should use todo lists for task tracking
4. Test with a conversation that involves multi-step work to see the capability in action

Your AI can then create and manage todo lists naturally during conversations, providing structured task tracking that makes complex work more organized and transparent.

Integration with Other Features

Session Todo Lists complement other ChatBotKit capabilities:

• Memory System: While todos handle temporary session tasks, memories persist long-term knowledge and preferences
• Secure Code Execution: Combine todo tracking with code execution for complex technical workflows where the agent tracks what code has been run and what remains
• Collaborative Spaces: Use todos to track file processing progress when working with Space contents

Session Todo Lists transform how AI agents handle complex, multi-step requests - bringing structure and visibility to work that would otherwise be opaque, and helping both agents and users stay aligned on progress toward goals.