Unlocking (even) more productivity with wider co-pilot support

The Kognic co-pilot is designed to maximize the value of auto-labels while maintaining high-quality standards. Unlike traditional auto-labeling approaches of just importing and using the label as-is, our co-pilot leverages imported predictions as prompts for our automation features. This creates a guided workflow that significantly reduces annotation time.

The process works as follows:

• Import your auto-labels into Kognic
• Identify the highest impact frames
• Use auto-labels as guidance for creating objects
• Enable one-click fixes with auto-adjustments
• Interpolate changes across all frames

With this approach, we've achieved up to 68% reduction in annotation time compared to manual methods, while maintaining annotation quality through efficient human feedback where needed most. Now imagine this being applied across all projects.

We're excited to share expanded auto-label capabilities, including:

• Relationship and geometry collection: Now you can leverage auto-labels for relationships and geometry collections.
• 3D Lanes
• 3D Points

These new features are already available now. For questions or support, feel free to reach out to your Kognic contact.