Grounded
AI skincare agent grounding demo

See how an AI agent grounds skincare answers before it recommends a routine.

Each scenario routes a request through the causal graph, ranks PubMed literature, plans AM / PM ingredients, and shortlists products with explicit confidence and safety context. Unsupported claims stay visible instead of becoming confident hallucinations.

Example skincare problems

Each scenario shows graph routing, routine ingredients, ranked AM/PM products, evidence, and caution context.

What makes this different

  • Conditions are routed through a causal knowledge graph, not just keyword search.
  • Ingredients are split into AM / PM with positive / negated context counts.
  • Evidence snippets link out to PubMed with short PMIDs.
  • Safety triggers and pregnancy filters are shown as context, not hidden.

Important note

This demo surfaces literature evidence, graph routing, and label / safety context alongside a product shortlist so users can make a more informed choice. It is not medical advice.