Genesys Cloud - Main

You not wrong.  The Enhanced TTS will only include the NTTS voices originally and the lower-costs voices will be added (maybe) will be added at a later date along with the newer languages.  

There seems some disconnect here on literature and options I am seeing in Architect, or I am missing something...

There are separate pricing tables for Enhanced TTS standard and Enhanced TTS Advanced.

The recent announcement points to using the Polly voices. In architect filtering on en-gb, polly I see 4 voice options, 2 male, 2 female.. all NTTS. The Advanced voices?

Where are the standard alternatives (in my case looking for en-gb) ?

Standard Genesys TTS is not the same thing, and only has en-gb, Ellene as option.

Thank you @Rebecca Harper for coming through as always.  I would really like to see this in the Resource Center since many may not find it here.  A great start to a great resource article if you ask me.  I'll probably blog on this soon.

Hi @Robert Wakefield-Carl, Michael Backus,

Thanks for your input here. We don't have an article explicitly listed which tags are supported but I have put together the list of related links we have to our material on the subject: 

https://developer.genesys.cloud/announcements/368880

https://help.mypurecloud.com/articles/migrate-to-enhanced-tts-offering-from-amazon-polly

https://help.mypurecloud.com/?p=379567

https://help.mypurecloud.com/announcements/deprecation-genesys-enhanced-tts-microsoft-azure-and-google-voices

https://help.mypurecloud.com/?p=368880

The most common SSML tags and their syntax for use in Genesys Cloud with Amazon Polly are:

1. Basic SSML Structure

<speak>

    This is the basic wrapper for all SSML content

</speak>

1. Break/Pause Tags

<speak>

    First sentence <break time="300ms"/> pause for 300 milliseconds

    Second sentence <break strength="medium"/> medium pause

    Third sentence <break strength="strong"/> strong pause

</speak>

1. Prosody (Control Speech)

<!-- Rate Control -->

<prosody rate="slow">This is spoken slowly</prosody>

<prosody rate="fast">This is spoken quickly</prosody>

<prosody rate="120%">Speed increased by 20%</prosody>

<!-- Pitch Control -->

<prosody pitch="high">Higher pitch</prosody>

<prosody pitch="low">Lower pitch</prosody>

<prosody pitch="+20%">20% higher pitch</prosody>

<!-- Volume Control -->

<prosody volume="loud">Louder volume</prosody>

<prosody volume="soft">Softer volume</prosody>

<prosody volume="+6dB">Increase volume by 6 decibels</prosody>

1. Emphasis

<speak>

    This is <emphasis level="strong">very important</emphasis>

    This is <emphasis level="moderate">somewhat important</emphasis>

    This is <emphasis level="reduced">less important</emphasis>

</speak>

1. Say-as (Pronunciation Control)

<!-- Numbers -->

<say-as interpret-as="cardinal">123</say-as>

<say-as interpret-as="ordinal">1st</say-as>

<!-- Dates -->

<say-as interpret-as="date" format="mdy">12-25-2024</say-as>

<!-- Phone Numbers -->

<say-as interpret-as="telephone">555-123-4567</say-as>

<!-- Characters -->

<say-as interpret-as="characters">ABC</say-as>

1. Phonetic Pronunciation

<phoneme alphabet="ipa" ph="təˈmeɪtoʊ">tomato</phoneme>

1. Language

<speak>

    This is in English

    <lang xml:lang="es-ES">Esto es en español</lang>

    Back to English

</speak>

1. Audio Mixing

<speak>

    <audio src="URL_to_audio_file.mp3">

        Fallback text if audio cannot be played

    </audio>

</speak>

If any of the above TAGs don't work for you I would recomment raising a support case so that we can look into it. 

@Michael Backus, according to Genesys, these are the supported ones:  Supported SSML tags - Amazon Polly.  Unfortunately, in my testing, I find the actually supported tags is much more limited.  I know they are working on supporting more of them, but at this point, it is hit of miss.  @Rebecca Harper, can you add a resource article (not and AWS link) that explains how to use these tags and which are fully supported?

The quote suggests that SSML support in Genesys is already fully available, but I’ve only been able to get it working in the most basic scenarios. I’m interested in learning more about the depth of AWS SSML support in Genesys Bot Flows, as it currently seems quite limited and there isn’t much documentation covering advanced usage.

Is Genesys planning any enhancements to allow users to more fully leverage the advanced features of AWS SSML, as referenced in the comment about “maximizing performance and taking advantage of AWS SSML features to the fullest”? Any insight would be appreciated.

 @Robert Wakefield-Carl

Hi @Neil Jones, thanks for your comment and apologies for the delayed response. It's great to hear that you are all set for the deprecation (which, incidently, has been delayed). 

You have a good question around failover plans. Depending on why kind of outage occurs, if it is with the TTS Engine for example, you could leverage another TTS Engine for the duration of the outage. It wiill be a degraded experience but better than a complete outage. 

We don't have individual SLAs for different functionality but, from my perspective, I'd say the voices would adhere to our platform SLAs as it is supported natively. I'd recommend talking to your CSM at Genesys to be sure, or if you have an incident. 

Hi, we've successfully migrated our TTS voices to a polly prefixed version, but i have a question around Genesys Enhanced TTS stability. Earlier this year (January 25) there was a TTS incident where voice playback was at double speed which impacted our voice bots.

We are looking to expand the use of TTS, i.e. migrating from traditional .wav prompt files played within inbound & in-queue flows, and as part of this migration I'm wondering if we need to build in a rollback process if a major Genesys Enhance TTS failure takes place.

Do Genesys provide any SLA information ?

Thanks, Neil

Hi @Kiarash Kazemi, thanks for your post. I agree this would be helpful but I don't believe we have a clear-cut list like that. Language Support will be updated under the Genesys Enhanced TTS expansion here: https://help.mypurecloud.com/articles/genesys-cloud-supported-languages/ and the list of voices will be somehwat limited to those from Amazon Polly. What will be available in the future will have dependency on what is available from Amazon. 

Hi, would you know if is there a full list of TTS of voices that will be supported for each language past the deprecation date? I think this would really be useful to share with clients that support multiple languages.

Hi @Amalric Villain, oh I see, I did misunderstand you. So this recording we made is on the deprecation of the voices being supported natively in Genesys Cloud via Genesys Enhanced TTS Engine. We're streamlining the voices we use in that service. Google supported voices are unchanged (well I can't vouch for other vendors but this recording doesn't relate to any changes to voice support that Google or MS are making). I hope that makes sense. 

My regards, 

My customer doesn't use the Genesys Enhanced integration, only the Google TTS integration. However, in the Google TTS integration, he has lost the same languages that you mentioned in your communication (which may make sense since both connectors should rely on the same Google APIs).

He does not use the custom voice feature, but I advised him to consider it (if the disappeared languages still exist at Google).

The idea of my post was to find out if the languages announced as removed were completely deleted at the source by Google or just from the standard of their API. I understand that the deleted voice in question may still exist and that they need to be integrated as custom voices.

If I misunderstood, please feel free to correct me

Hi @Amalric Villain, thanks for your comment!

Hopefully I understand what you are asking here; yes, all the Google and Azure Voices will be removed from the Genesys Enhanced TTS Engine for selection and the options are to instead use their respective integrations to continue to use those voices. 

We recorded the Googe Integration process. Then to use custom voices within that integration you should follow this: https://help.mypurecloud.com/articles/configure-google-cloud-tts-custom-voices/ 

If your customer is using the custom voices then they should already be using the integration and not the GC Enhanced TTS?

Hello,
Since the topic is quite similar, I would like to respond to the announcement: https://fr-help.mypurecloud.com/announcements/deprecation-genesys-enhanced-tts-google-european-voices/

The announcement seems to correspond only to the Genesys Enhanced TTS integration; however, I have a client who has just noticed (rather late) that the voices mentioned in this announcement also seem to apply to the Google TTS integration you offer elsewhere (https://appfoundry.genesys.com/filter/genesyscloud/listing/1938bab8-e163-46f9-ba09-5359a3d8bd16).

 Is this also the case for this integration? (and therefore, was the communication not complete) or does the client need to switch to custom voices (https://community.genesys.com/viewdocument/custom-voices-support-for-the-googl?CommunityKey=bab95e9c-6bbe-4a13-8ade-8ec0faf733d4&tab=librarydocuments) to keep their old voices?

Thank you in advance for your response.

Discussing this topic with product, so evidently this comment is not a commitment statement, and there is a plan to support Polly Generative voices via Enhanced TTS later this year (on top of the existing support of the Neural Polly voices). This will help address the need to have more varied gender voice options and languages. Until then you will have to use the integration options. On that topic we have already recorded one: Google Cloud TTS Integration

Hi @Hannah Mitchell,

Thanks for commenting. As far as I am aware, there is no pricing changes planned, at present. Genesys provides the Standard TTS at no cost (however, I think you'll find this option not suitable for good conversational sounding voice quality) and the price of Enhanced TTS (Polly voices) is unchanged from the current pricing for Enhanced TTS. 

Currently I don't see different pricing options for access to other voices in the Enhanced TTS space. Your only option, to open the range of voices, it to use one the the TTS integrations which wears a smaller consumption-based cost on the Genesys side and a cost for the third party vendor of your choice. 

Hi - how will this affect standard vs. advanced pricing for TTS for en-GB Amazon Polly voices (if there is standard vs. advanced optionality)?

Yes, I have addressed this internally and with some clients that the reasoning is not just political, but the ability to maximize performance and take advantage of the AWS SSML features to the fullest.  Yes, this change is a huge hassle for most of our clients and without automation, could mean significant costs for them.  I also hope that AWS keeps up with the TTS market and don't let it whither as they have Lex.  

Hi Robert,

Yes, the multilingual TTS functionality is something that would be amazing to have and I, personally, live in hope AWS will have a competitive option in the future. I can't speak for product rational here, but I know that security and stability are the top priorities of Genesys, and it seems like a primary driver for this decision.

Your comment around the manual overhead, that is exactly the concern our teams can talk to a customer about. Every customer’s environment and needs will be different, so I am hesitant to make recommendations on this thread that might not be the right approach for all. We have some methods, not native to Genesys Cloud, that could be leveraged, or some PS packages that we can offer. So it’s best to discuss this at an individual level right now.

The other issue is that MS and Google now offer multilingual TTS so you can have the same voice across multiple languages.  AWS does not offer this and those using Enhanced TTS will miss out on this feature as well.  

Have you examined the effort required to swap voices in flows?  We have not found a solution that requires manually opening EACH and EVERY flow, changing the settings and republishing.  That is about 5 minutes per flow - now imaging if you have 2500 flows!

Hi @Robert Wakefield-Carl, 

Thank you for kicking off this conversation and forgive the delayed response. We wanted to cross check our understanding before replying to what we anticipate will be a hot topic for some time. 

You are right, there are no further exceptions planned and languages that do not have a Polly Language available (with those few exceptions) will have to revert to BYO TTS Engine. We are waiting on the AWS roadmap for their Polly Languages for the coming year.

You will see an upcoming integration that should reduce some of the difficulties of BYO TTS engine: https://help.mypurecloud.com/articles/activate-and-configure-the-genesys-tts-connector-integration/ or the existing: https://help.mypurecloud.com/articles/install-the-microsoft-azure-cognitive-services-tts-integration/ for the Azure voices. 

We are encouraging customers to reach out to their partners or Genesys representatives, if they are impacted by this deprecation. We are putting together some options to minimize the expected manual overhead or even leverage our professional services if desired.

There is an exception for Thai and Basque, but what about all the other languages not supported like Russian, Haitian, and many others that don't have Neural voices?