I'm trying to unravel some conflicting information on STT and TTS pricing

Genesys enhanced TTS is free to use in Bot flows as per

https://help.mypurecloud.com/articles/genesys-enhanced-tts-pricing/

But this article indicates it is only for the google voices, and not the azure ones? But the pricing article doesn't mention that.

https://help.mypurecloud.com/faqs/how-does-genesys-bill-text-to-speech-tts-with-google-cloud-dialogflow/

So is the FAQ right or the pricing article?

And then what about STT, if I use a language that is otherwise not available without enhanced TTS and extended voice transcription which I'm not using, what does it use for the STT component to do the intent and slot filling? And what cost would that have if any?

If anyone has some documentation or info that can help me make sense of all that, it would be appreciated.

Hi Anton,

I recently went through an excercise where I did a deep dive on the usage of STT withinside of a Genesys Bot Flow. Here would be my responses to your comments:

This article you reference is specifically calling out the scenario where: we are google voices, through the selection of the Genesys Enhanced TTS provider, and we are calling a Google Cloud Dialogflow CX Bot. It is not related to using Genesys Enhanced TTS inside of the Genesys Bot flows. Using Google Voices through Enhanced TTS inside of a Bot Flow is free, as stated on the pricing page.

Extended Voice Transcription Services do not apply to Bots. That is only applicable to the Speech and Text Analytics component, from my perspective.

STT is best leveraged when you think "hey, I don't like how the Genesys native ASR/STT is performing".  Maybe there is something funky going on with how genesys is transcribing words. So you opt for an STT engine (only Microsoft Azure Cog. services or Google STT are available).  What is kind of misleading, at least in my opinion.. is that STT engines are limited to the languages that Bot Flows support. Meaning: you can't use an STT engine to expand Bot Flows language support. My understanding is that even though we are using a 3rd Party STT, we would still leverage Genesys' NLU (Intent recognition) capabilities.

Pricing for the STT Engines can be foud here: https://help.mypurecloud.com/articles/bring-your-own-technology-services-model-per-turn-minute-rates-current-model/.

Does that help?

I'm trying to unravel some conflicting information on STT and TTS pricing

Genesys enhanced TTS is free to use in Bot flows as per

https://help.mypurecloud.com/articles/genesys-enhanced-tts-pricing/

But this article indicates it is only for the google voices, and not the azure ones? But the pricing article doesn't mention that.

https://help.mypurecloud.com/faqs/how-does-genesys-bill-text-to-speech-tts-with-google-cloud-dialogflow/

So is the FAQ right or the pricing article?

And then what about STT, if I use a language that is otherwise not available without enhanced TTS and extended voice transcription which I'm not using, what does it use for the STT component to do the intent and slot filling? And what cost would that have if any?

If anyone has some documentation or info that can help me make sense of all that, it would be appreciated.

From what it sounds like, and this why it is confusing, when Genesys says TTS they mean language model, so if I have a language set on a flow, bot or inbound, that will handle both TTS and the understanding (which my understand is, that is done by converting it to text so STT and then doing the logical checks on the text values - maybe I'm wrong here and that is why its confusing).

There is also the extend voice transcription for STT - https://help.mypurecloud.com/faqs/voice-transcription-how-much-does-extended-voice-transcription-services-cost/ Which is cheaper than the BYOT Rate E rates. Which includes some Azure cognitive services language options. 

Hi Anton,

I recently went through an excercise where I did a deep dive on the usage of STT withinside of a Genesys Bot Flow. Here would be my responses to your comments:

This article you reference is specifically calling out the scenario where: we are google voices, through the selection of the Genesys Enhanced TTS provider, and we are calling a Google Cloud Dialogflow CX Bot. It is not related to using Genesys Enhanced TTS inside of the Genesys Bot flows. Using Google Voices through Enhanced TTS inside of a Bot Flow is free, as stated on the pricing page.

Extended Voice Transcription Services do not apply to Bots. That is only applicable to the Speech and Text Analytics component, from my perspective.

STT is best leveraged when you think "hey, I don't like how the Genesys native ASR/STT is performing".  Maybe there is something funky going on with how genesys is transcribing words. So you opt for an STT engine (only Microsoft Azure Cog. services or Google STT are available).  What is kind of misleading, at least in my opinion.. is that STT engines are limited to the languages that Bot Flows support. Meaning: you can't use an STT engine to expand Bot Flows language support. My understanding is that even though we are using a 3rd Party STT, we would still leverage Genesys' NLU (Intent recognition) capabilities.

Pricing for the STT Engines can be foud here: https://help.mypurecloud.com/articles/bring-your-own-technology-services-model-per-turn-minute-rates-current-model/.

Does that help?

I'm trying to unravel some conflicting information on STT and TTS pricing

Genesys enhanced TTS is free to use in Bot flows as per

https://help.mypurecloud.com/articles/genesys-enhanced-tts-pricing/

But this article indicates it is only for the google voices, and not the azure ones? But the pricing article doesn't mention that.

https://help.mypurecloud.com/faqs/how-does-genesys-bill-text-to-speech-tts-with-google-cloud-dialogflow/

So is the FAQ right or the pricing article?

And then what about STT, if I use a language that is otherwise not available without enhanced TTS and extended voice transcription which I'm not using, what does it use for the STT component to do the intent and slot filling? And what cost would that have if any?

If anyone has some documentation or info that can help me make sense of all that, it would be appreciated.

Hi Anton,

First I'll say that Peter's explanation is good. Thanks Peter! Genesys Enhanced TTS is included in the Bot Flows pricing for all of the Genesys Enhanced TTS Voices. The second link does not apply to Bot Flows, it's only talking about Google Dialogflow Bots.

There are 3 components involved for audio Bot Flows: Speech-To-Text, Natural Language Understanding, and Text-To-Speech. The STT and NLU provide the "understanding" that you mentioned above.

For Bot Flows, the TTS Engine & Voice that get used are the same as the Inbound Call Flow which called the Bot Flow. This way the TTS Engine & Voice is consistent between the Inbound Call Flow actions and the Bot Flow actions. In the Inbound Call Flow, you'll select a different TTS Voice for each of the languages that your Inbound Call Flow supports. Those TTS Voices could all be from the same TTS Engine or you can mix voices from different TTS Engines.

So that means for Bot Flows the only configuration is the STT engine. By default, it is unset and Genesys will provide one. As Peter said, if there is a reason the provided one isn't working, we have the BYOT STT options to allow for customization. Like the TTS Voice, this setting is per language, so you could use the Genesys provided STT Engine for one language and a BYOT STT engine for a different language. Peter is right that despite BYOT STT, the Bot Flow is still limited to the languages we support due to the need for an NLU model and other ancillary resources for the language. We are continually working to expand our language coverage.

The EVTS does not apply for Bot Flows. That transcription is separate from what Bot Flows use.

Hopefully that clears everything up.

From what it sounds like, and this why it is confusing, when Genesys says TTS they mean language model, so if I have a language set on a flow, bot or inbound, that will handle both TTS and the understanding (which my understand is, that is done by converting it to text so STT and then doing the logical checks on the text values - maybe I'm wrong here and that is why its confusing).

There is also the extend voice transcription for STT - https://help.mypurecloud.com/faqs/voice-transcription-how-much-does-extended-voice-transcription-services-cost/ Which is cheaper than the BYOT Rate E rates. Which includes some Azure cognitive services language options. 

Hi Anton,

I recently went through an excercise where I did a deep dive on the usage of STT withinside of a Genesys Bot Flow. Here would be my responses to your comments:

This article you reference is specifically calling out the scenario where: we are google voices, through the selection of the Genesys Enhanced TTS provider, and we are calling a Google Cloud Dialogflow CX Bot. It is not related to using Genesys Enhanced TTS inside of the Genesys Bot flows. Using Google Voices through Enhanced TTS inside of a Bot Flow is free, as stated on the pricing page.

Extended Voice Transcription Services do not apply to Bots. That is only applicable to the Speech and Text Analytics component, from my perspective.

STT is best leveraged when you think "hey, I don't like how the Genesys native ASR/STT is performing".  Maybe there is something funky going on with how genesys is transcribing words. So you opt for an STT engine (only Microsoft Azure Cog. services or Google STT are available).  What is kind of misleading, at least in my opinion.. is that STT engines are limited to the languages that Bot Flows support. Meaning: you can't use an STT engine to expand Bot Flows language support. My understanding is that even though we are using a 3rd Party STT, we would still leverage Genesys' NLU (Intent recognition) capabilities.

Pricing for the STT Engines can be foud here: https://help.mypurecloud.com/articles/bring-your-own-technology-services-model-per-turn-minute-rates-current-model/.

Does that help?

I'm trying to unravel some conflicting information on STT and TTS pricing

Genesys enhanced TTS is free to use in Bot flows as per

https://help.mypurecloud.com/articles/genesys-enhanced-tts-pricing/

But this article indicates it is only for the google voices, and not the azure ones? But the pricing article doesn't mention that.

https://help.mypurecloud.com/faqs/how-does-genesys-bill-text-to-speech-tts-with-google-cloud-dialogflow/

So is the FAQ right or the pricing article?

And then what about STT, if I use a language that is otherwise not available without enhanced TTS and extended voice transcription which I'm not using, what does it use for the STT component to do the intent and slot filling? And what cost would that have if any?

If anyone has some documentation or info that can help me make sense of all that, it would be appreciated.

Thanks @Brad Wehmeier

So just want to be clear

The TTS we set in the inbound flow, has no bearing at all on the NLU/STT component for the bot. So if we are using Genesys Enhanced TTS to use an Azure voice the NLU/STT the bot uses will be the Genesys engine, not Azure, not Google. So even with both enhanced TTS and EVTS if we wanted a different engine in bot we would need to go down the BYOT path? 

When we select the default supported language in the bot, we are setting the language for the Genesys Engine NLU/STT, and this is a different engine used in any other flow with different languages supported?

Hi Anton,

First I'll say that Peter's explanation is good. Thanks Peter! Genesys Enhanced TTS is included in the Bot Flows pricing for all of the Genesys Enhanced TTS Voices. The second link does not apply to Bot Flows, it's only talking about Google Dialogflow Bots.

There are 3 components involved for audio Bot Flows: Speech-To-Text, Natural Language Understanding, and Text-To-Speech. The STT and NLU provide the "understanding" that you mentioned above.

For Bot Flows, the TTS Engine & Voice that get used are the same as the Inbound Call Flow which called the Bot Flow. This way the TTS Engine & Voice is consistent between the Inbound Call Flow actions and the Bot Flow actions. In the Inbound Call Flow, you'll select a different TTS Voice for each of the languages that your Inbound Call Flow supports. Those TTS Voices could all be from the same TTS Engine or you can mix voices from different TTS Engines.

So that means for Bot Flows the only configuration is the STT engine. By default, it is unset and Genesys will provide one. As Peter said, if there is a reason the provided one isn't working, we have the BYOT STT options to allow for customization. Like the TTS Voice, this setting is per language, so you could use the Genesys provided STT Engine for one language and a BYOT STT engine for a different language. Peter is right that despite BYOT STT, the Bot Flow is still limited to the languages we support due to the need for an NLU model and other ancillary resources for the language. We are continually working to expand our language coverage.

The EVTS does not apply for Bot Flows. That transcription is separate from what Bot Flows use.

Hopefully that clears everything up.

From what it sounds like, and this why it is confusing, when Genesys says TTS they mean language model, so if I have a language set on a flow, bot or inbound, that will handle both TTS and the understanding (which my understand is, that is done by converting it to text so STT and then doing the logical checks on the text values - maybe I'm wrong here and that is why its confusing).

There is also the extend voice transcription for STT - https://help.mypurecloud.com/faqs/voice-transcription-how-much-does-extended-voice-transcription-services-cost/ Which is cheaper than the BYOT Rate E rates. Which includes some Azure cognitive services language options. 

Hi Anton,

I recently went through an excercise where I did a deep dive on the usage of STT withinside of a Genesys Bot Flow. Here would be my responses to your comments:

This article you reference is specifically calling out the scenario where: we are google voices, through the selection of the Genesys Enhanced TTS provider, and we are calling a Google Cloud Dialogflow CX Bot. It is not related to using Genesys Enhanced TTS inside of the Genesys Bot flows. Using Google Voices through Enhanced TTS inside of a Bot Flow is free, as stated on the pricing page.

Extended Voice Transcription Services do not apply to Bots. That is only applicable to the Speech and Text Analytics component, from my perspective.

STT is best leveraged when you think "hey, I don't like how the Genesys native ASR/STT is performing".  Maybe there is something funky going on with how genesys is transcribing words. So you opt for an STT engine (only Microsoft Azure Cog. services or Google STT are available).  What is kind of misleading, at least in my opinion.. is that STT engines are limited to the languages that Bot Flows support. Meaning: you can't use an STT engine to expand Bot Flows language support. My understanding is that even though we are using a 3rd Party STT, we would still leverage Genesys' NLU (Intent recognition) capabilities.

Pricing for the STT Engines can be foud here: https://help.mypurecloud.com/articles/bring-your-own-technology-services-model-per-turn-minute-rates-current-model/.

Does that help?

I'm trying to unravel some conflicting information on STT and TTS pricing

Genesys enhanced TTS is free to use in Bot flows as per

https://help.mypurecloud.com/articles/genesys-enhanced-tts-pricing/

But this article indicates it is only for the google voices, and not the azure ones? But the pricing article doesn't mention that.

https://help.mypurecloud.com/faqs/how-does-genesys-bill-text-to-speech-tts-with-google-cloud-dialogflow/

So is the FAQ right or the pricing article?

And then what about STT, if I use a language that is otherwise not available without enhanced TTS and extended voice transcription which I'm not using, what does it use for the STT component to do the intent and slot filling? And what cost would that have if any?

If anyone has some documentation or info that can help me make sense of all that, it would be appreciated.