What Amplitude shipped
Amplitude shipped Agent Analytics to solve a modern product problem: measuring the actual ROI of AI features. Standard analytics treat all AI interactions as generic engagement, obscuring whether an agent hallucinated, stalled, or actually helped the user.
This launch film introduces the solution by visualizing the bridge between LLM observability and traditional product analytics, proving that AI quality and business outcomes can exist on the same identity graph.
How the motion works
The film has to visualize a conceptual bridge—connecting the black box of an AI agent to hard product metrics. It relies on kinetic-typography to establish the problem, using rapid text reveals to list out the invisible failures of AI before snapping to the interface.
When transitioning to the product, the motion uses a match-cut to move from abstract data nodes directly into the Amplitude identity graph. The camera pushes in on specific user journeys, using a punch-in technique to isolate the exact moment an AI session fails and how it impacts the subsequent funnel.
Pacing is dictated by the complexity of the data. The edit respects the cognitive load of the viewer, utilizing generous hold-time on the charts to let the audience register the connection between agent quality and retention. Easing is sharp and exponential, keeping the technical subject matter feeling responsive and precise.
What to steal from it
- Use kinetic typography to clearly define the user problem before introducing the interface.
- Match-cut abstract concepts directly into UI components to ground theoretical features in reality.
- Ensure sufficient hold-time when displaying complex data visualizations or identity graphs.
- Drive camera movements with purpose, using punch-ins to direct the viewer's eye to critical metrics.
Want a video like this for your product?
Impractical cuts one from a single prompt — the same motion craft, in about twenty minutes.
