The Genesys TTS Deadlines Have Started Passing | Here Is Your BYOT-A Migration Plan

If your contact center runs on Genesys Cloud and uses Enhanced Text-to-Speech, two deadlines have already passed and a third is coming. Genesys ended support for Standard Enhanced TTS on August 5, 2026. The action deadline for Google Standard TTS voices followed on August 10, 2026. If you use Google WaveNet, Google Neural, or Microsoft Azure Neural voices, your deadline is March 2027.

If you missed the August deadlines, your prompts may already be failing or falling back to unsupported behavior. If you are on the March 2027 timeline, you have a decision to make, and it is a bigger decision than most teams realize.

This guide covers what BYOT-A migration actually involves, what it costs, and the alternative that removes deprecation risk from your call flows permanently.

What Happened

Genesys is removing native support for select Google and Microsoft TTS voices in Genesys Cloud. After each end of support date, those voices are treated as third party integrations rather than native features. If no action is taken, the affected text-to-speech functionality may stop working depending on your configuration.

The rollout has three key dates:

• August 5, 2026: End of support for Standard Enhanced TTS
• August 10, 2026: Action deadline for Google Standard TTS voices
• March 2027: Action deadline for Google WaveNet, Google Neural, and Microsoft Azure Neural voices

Genesys has stated the goal is to streamline its partner portfolio and keep TTS processing contained within AWS. That is a reasonable business decision for Genesys. It also means every contact center that built call flows on Google or Microsoft voices now has to rebuild, repay, or replace.

What Is BYOT-A?

BYOT-A stands for Bring Your Own TTS, Rate A. It is the billing model Genesys uses for third party text-to-speech integrations accessed through the AppFoundry marketplace.

Under BYOT-A, you contract directly with Google or Microsoft for the voices, install the integration from AppFoundry, configure it in your org, and pay Genesys per character generated. As of January 1, 2026, the BYOT-A rate is $4 per million characters, billed in one million character increments rounded up.

In plain terms, BYOT-A means:

1. A new vendor contract with Google or Microsoft
2. A new integration to install, configure, and maintain
3. A new line item on your Genesys bill that scales with call volume
4. The same synthetic voices your callers were hearing before

You pay more to keep what you had.

Your Three Options

Option 1: Migrate to AWS Neural Polly voices

Genesys recommends this path because Polly voices remain natively supported. It is the lowest friction option inside the platform. The tradeoff is that your callers get a different synthetic voice than the one you tuned your flows around, and every prompt needs to be reviewed for pronunciation, pacing, and SSML behavior under the new engine. You are still exposed to the next deprecation cycle, whenever it comes.

Option 2: Migrate to BYOT-A

If your organization is committed to Google or Microsoft voices, BYOT-A keeps them available. Plan for the direct vendor contract, the AppFoundry integration work, the per character billing, and ongoing maintenance whenever Google or Microsoft update their voice APIs. This is the highest cost, highest maintenance path.

Option 3: Replace synthetic prompts with professionally recorded human audio

This is the option Genesys documentation does not spend much time on, and it is the only one that ends the cycle. Recorded prompts are audio files. They do not depend on a TTS engine, an API contract, a per character rate, or a vendor roadmap. Once a prompt is recorded and loaded into Architect, no deprecation notice can ever break it.

Recorded prompts also solve the problem TTS never has: callers can hear the difference. A professional voice with correct pronunciation, natural pacing, and consistent brand tone outperforms synthetic speech on caller experience, and it is one less system that can fail during an outage or migration.

What a Migration to Recorded Audio Looks Like

At COHM we have been producing IVR and contact center audio for over 40 years, including for Genesys Cloud environments. A typical TTS to recorded audio migration follows five steps:

1. Prompt audit. We export and inventory every TTS prompt in your Architect flows, including IVR menus, queue announcements, voicemail greetings, and bot flow audio.
2. Script cleanup. TTS scripts are written for machines. We rewrite them for a human voice, fixing awkward phrasing, pronunciation traps, and inconsistent terminology across departments.
3. Recording. Professional voice talent records every prompt with consistent tone, pacing, and pronunciation standards. Multilingual programs are handled with native speakers per language.
4. Genesys Cloud formatting. Files are delivered in the exact formats Genesys Cloud requires, named to match your flow structure, ready to load into Architect without conversion work.
5. Updates on demand. When menus change, new prompts are recorded by the same voice with the same standards, so your system never sounds patched together.

For dynamic content that genuinely requires synthesis, such as reading back account numbers or wait times, a hybrid approach works well: recorded prompts for everything static, with a minimal TTS footprint only where variables demand it. That shrinks your BYOT-A exposure to a fraction of your audio.

What This Costs Compared to BYOT-A

BYOT-A is a recurring cost that scales with call volume forever. Recorded audio is a one time production cost per prompt, with small incremental costs only when content changes. For most contact centers, the static prompts that make up the majority of caller-facing audio pay for themselves quickly against per character billing, and the caller experience improvement comes free.

If you want a specific comparison for your call volume and prompt inventory, we will run the numbers with you before you commit to anything.

Frequently Asked Questions

The Bottom Line

The Genesys TTS deprecation is not a one time event. It is the latest reminder that synthetic voice infrastructure sits on someone else’s roadmap. You can migrate to Polly and wait for the next notice, pay BYOT-A rates to keep what you had, or move your prompts to recorded human audio and never read a deprecation announcement again.

If you want help auditing your prompts or planning the migration, contact COHM. We have been keeping enterprise phone systems sounding professional since 1983.