Generative UI Template
Turn imagination into interface.
Generative UI Templates turn AI results into interactive experiences. Instead of answering with text alone, an application responds with charts, tables, forms and actions selected for the task.
Interfaces that answer the question
Generative UI lets an AI application respond with useful interface elements alongside text. A request such as “Compare these facilities” can return a comparison table, a trend chart, filters and buttons for investigating individual facilities.
The model calls a tool, the tool returns structured data, and the application renders that data through a registered component. The agent chooses from components you have approved and supplies validated data and parameters, so you stay in control of presentation, interactions and rendering behavior.
What a template defines
| Template element | Purpose |
|---|---|
| Component type | Chooses a chart, table, form, card or another supported view |
| Data contract | Specifies the fields and types the component accepts |
| Presentation settings | Defines labels, units, formatting and layout |
| Interaction contract | Defines what filtering, selection or submission does |
| Loading and error states | Handles incomplete results and failures |
| Theme and accessibility | Keeps experiences consistent across applications |
Example: an equipment performance comparison
- A user asks for an equipment performance comparison.
- An agent retrieves the relevant measurements.
- An analysis function produces structured comparison data.
- The application selects a comparison template.
- The user filters the results or opens a detailed investigation.
The same interface contract serves results from either Agent Run framework. The platform translates framework-specific output into the agreed component data format.
Template library
- KPI summaries and trend charts.
- Search results and comparison tables.
- Document extraction review forms.
- Approval panels.
- Investigation timelines.
- Financial research dashboards.
Unlike a conventional template library, the application selects and fills the interface according to the user's task and the results available.
Make AI answers actionable: inspect the evidence, adjust inputs, compare results and continue the workflow from the interface.