Skip to content

🔭 Product Roadmap

This page outlines the upcoming features and long-term vision for Airflow Copilot. While the current version offers powerful orchestration capabilities via Microsoft Teams, several enhancements are actively being explored to make the assistant more intelligent, proactive, and versatile.

🧠 Planned Enhancements

1. 🔧 Tooling Optimization

Improve the performance and accuracy of existing tools used by the AI agent—especially around error handling, fallback mechanisms, and dynamic Airflow version support.

2. 🗄️ Redis Support as Copilot Backend

Introduce Redis as an optional backend for storing Copilot state, checkpoints, and intermediate responses. This addition aims to:

  • Enhance performance for high-concurrency environments.
  • Reduce latency during message summarization and session context recall.
  • Support distributed deployments with better scalability.
  • Serve as an alternative to PostgreSQL for ephemeral state tracking.

3. 📡 Multi-Channel Communication

Expand Copilot’s availability to support other messaging platforms beyond Microsoft Teams (e.g., Slack, WhatsApp, Webchat). This ensures greater accessibility for users across teams and infrastructures.

4. 🔔 Proactive Monitoring (User-Driven)

Introduce intent-based monitoring. Example:

“Copilot, monitor my_dag tonight and notify me if it fails.”

Copilot will intelligently track execution outcomes and proactively notify the user based on specific instructions.

5. ✉️ Email Notification Integration

Enable direct email support for DAG status updates, alerts, and action confirmations. This allows users to receive updates even outside their chat interface.

6. 🔑 Allow per user LLM Key and Model

Currently the LLM Key is centralised that can be bottelneck for some case so plan it to provide deployment as BYOK (Bring you own Key).

💡 Have a Feature Request?

I'd love to hear from you! Feel free to open a GitHub issue or suggest improvements via your preferred channel.

ℹ️ Note: Roadmap items are subject to change based on user feedback, platform limitations, and emerging use cases. Stay tuned for updates!