When Lexicon matches an intent at runtime it looks for an action associated with that intent and calls its Process method with a LexiconRuntimeResult. The LexiconRuntimeResult contains the following information:

Utterance	The phrase that was uttered by the user
Intent	The intent that matched this utterance
Confidence	Confidence score for the intent match, a value between 0 and 1
EntityMatches	The EntityMatches that were found in this utterance
SpeechResult	The full speech-to-text result for this utterance

LexiconRuntimeResult also contains a few helpful methods:

GetEntityMatch(string entityName)	Returns the first entity match for the given entity name
GetEntityMatches(string entityName)	Returns all the entity matches for the given entity name
GetEntityBefore(string entityName, LexiconEntityMatch targetMatch)	Returns the entity with the given entity name that appears before the target match
GetEntityAfter(string entityName, LexiconEntityMatch targetMatch)	Returns the entity with the given entity name that appears after the target match

Use these methods to parse an utterance that contains multiple parts. For example, if a user says “Create a red cube here and a blue sphere here”, we can use the following approach:

1. Use GetEntityMatches to find all the primitives in the utterance
2. Loop over the primitive matches
   • Use GetEntityBefore to find the color for the current primitive
   • Use GetEntityAfter to find the position for the current primitive