Misunderstanding on pricing for Dialogflow CX

[a_train]

I am looking at pricing of Dialogflow but can't understand a detail.

Can someone please explain how we can have 50% of intent-based queries ? Aren't intents always checked on queries and when none is found then generative becomes an option ?


Use case from https://cloud.google.com/generative-ai-app-builder/pricing 

Example of mixing intent-based and generative answers: Customer Support on Website

• Intent-based, deterministic flows are used for some customer-specific questions (Billing, Payment)
• Generative answers used to handle questions where no intent defined but answer is on the website

Assumptions:

• 100M annual website visits
• 5% of visitors use the chatbot, 3 queries per use, 50% generative, 50% intent-based

SKU Volume Rate Total List PriceDialogflow CX Chat (intent-based)Dialogflow CX Enterprise Chat (generative)Data Indexed

7.5M annual 100M*5% use*3 queries*50% intent-based	* $7.00 / 1k queries =	$52,500
7.5M annual 100M*5% use*3 queries*50% generative	* $12.00 / 1k queries =	$90,000
50GB annual 10GB free so 40GB used in price calc	* $5.00 / GB * 12 month =	$2,400

[ibaui]

Hi @a_train,

Welcome to Google Cloud Community!

It seems like you are facing confusion regarding how Dialogflow can handle a 50/50 mix of intent-based and generative queries. You're also seeking clarification on the pricing model and the use case example given in the Google Cloud documentation for Dialogflow.

The documentation says that you can have 50% intent-based queries. This doesn't mean that Dialogflow only checks for intent-based queries first. It processes all the queries, but it charges based on the type of response that is used.

Here's a breakdown of how it works:

• If Dialogflow can match a user query to a specific intent, it provides an intent-based response. This is the most cost-effective way to handle queries.
• If Dialogflow can't match a user query to an intent, it might be able to generate a response using its generative capabilities. This is a more expensive way to handle queries, but it can still be helpful for providing users with some information.

In the example provided in the documentation, they assume that 50% of the queries will match an intent, while the other 50% will require a generative response. However, the actual distribution will depend on your specific use case and how you design the conversation flow.

Additionally, I recommend reviewing the official Dialogflow pricing page to obtain accurate pricing information for your specific scenario.

Was this helpful? If so, please accept this answer as “Solution”. If you need additional assistance, reply here within 2 business days and I’ll be happy to help.

[a_train]

Thanks for your explanation.  However I still have questions.

"If Dialogflow can't match a user query to an intent, it might be able to generate a response using its generative capabilities"

That's what I don't understand. 

In order to do that, it need to check for intents and according to https://cloud.google.com/dialogflow/pricing we pay for "Text
(includes all DetectIntent, StreamingDetectIntent, and FulfillIntent requests that do not contain audio)" . Because of DetectIntent, all queries should have this intent-based fee.

Or should I understand that if DetectIntent fails, I am not charged ? Then, the documentation may be clarified on this point.

PS:
There's also the case where I can match an Intent with a generative answer so that my intent-based answer is generative and that's not well explained. Will I pay for the detectIntent + generative (or only one)  ?