While routing based on VF intents/utterance examples is (in my experience so far) the least precise way of doing it, it doesn't cost any tokens and is faster. Using a LLM (either through the settings on the Intents page or setting it up custom in the flow) with a strong prompt will be more accurate in many scenarios, but it will cost tokens and take longer. So it's about tradeoffs and what you're willing to accept for your use case.
While routing based on VF intents/utterance examples is (in my experience so far) the least precise way of doing it, it doesn't cost any tokens and is faster.
Using a LLM (either through the settings on the Intents page or setting it up custom in the flow) with a strong prompt will be more accurate in many scenarios, but it will cost tokens and take longer.
So it's about tradeoffs and what you're willing to accept for your use case.
Thank you very much, my brother!!
Thanks G
Thank you very much, was really struggling with intents
thanks buddy
How can I get the Ai prompt ? Could ypu write it here? Great Video!!
Using ai model in voiceflow is a great idea, but don't forget about the tokens consumption