Dynamically Generated Experiences
DGE uses LLMs and RAG grounded in published Sitefinity content to assemble intent-based page sections in real time. Explore the FAQ below for more information.
Core Concept and Functionality
Technical Architecture and Requirements
- What does the high-level architecture look like?
- How does DGE communicate with the chosen LLM?
- Which LLMs are supported?
- How does Sitefinity fetch content for generation?
- Does DGE work only with Sitefinity content or can it retrieve material from external systems such as CRM, CDP, knowledge bases or documentation sources?
- Can I restrict certain content such as drafts, internal docs or sensitive items?
- How is security handled between Sitefinity and external models?
Data, Training and Content Use
- What content can be used to ground or train the DGE output?
- How is the grounding data indexed or prepared?
- Can different data sources carry different weights or rules?
- How do I update the training content when something changes on the site?
- How is hallucination reduced?
- How is bias handled?
- How does DGE avoid exposing content that should remain private?
Performance and Scale
- What is the expected render time for a dynamically generated experience?
- How does DGE behave under heavy traffic?
- Can the service scale for many concurrent users?
- What architectural choices speed up generation?
- Can I prebuild a page for faster first views using query strings or prewarming techniques?
- What happens on timeout or fallback scenarios?