Speech
Speech is the base WinnerWare module for spoken experiences. It provides the shared foundation used when the product needs to play voice guidance, work with transcript records, or support speech-enabled workflows.
Overview
This module is the common layer behind speech features in WinnerWare. It supports the voice and transcript capabilities that provider-specific modules build on later.
This is the part of the solution that helps teams:
- choose spoken voices for guided experiences
- keep transcript records available for follow-up work
- support speech-based workflows and downstream reporting
Key Features
- Provides the shared speech foundation used by the rest of the speech features.
- Supports voice selection for speech-enabled experiences.
- Supports providers that turn speech into text, including live (realtime) speech-to-text through ElevenLabs.
- Supports transcript records and transcript task tracking.
- Supports secured access to redacted recordings when those are available.
- Supports workflow-based follow-up when transcript processing is completed.
- Supports search and indexing of transcript content when those capabilities are turned on elsewhere in WinnerWare.
How It Works
Shared speech foundation
This module is the common base for speech-enabled features in WinnerWare. It does not represent one provider by itself. Instead, it gives the rest of the speech area a shared place to work from.
That is why it is usually paired with a provider module such as Speech Azure or Speech ElevenLabs.
Voice-driven experiences
The module supports voice selection for speech-enabled content and guided experiences. In practice, this helps teams keep spoken guidance consistent across the product instead of picking voices ad hoc.
Transcript-driven experiences
When transcript features are in use, the module keeps transcript records and transcript processing tasks available for the rest of WinnerWare. This is what allows speech-based content to continue into follow-up actions such as review, reporting, or workflow handling.
Configuration
Choose a speech provider
The base Speech module works best as the common layer under a provider:
- use Speech Azure when Azure should handle speech and transcription, including Speech Insights call recordings
- use Speech ElevenLabs when ElevenLabs should provide the spoken voice experience, speech-to-text, or live (realtime) speech-to-text
Enable the right speech features
| Admin display name | Use when |
|---|---|
| Speech | You need the shared speech foundation. Other speech modules depend on it. |
| Speech Transcriptions | You need transcript records, for example for Speech Insights. |
| Speech Service (Azure) | Azure should handle speech and transcription. See Speech Azure. |
| Speech Services (Eleven Labs) | ElevenLabs should provide voices, text-to-speech, speech-to-text, or realtime speech-to-text. See Speech ElevenLabs. |
Usage
Build a speech-enabled experience
A typical WinnerWare speech rollout starts with the shared Speech feature, then adds the provider module that will handle the actual voice or transcription service.
That pattern helps teams keep one consistent speech foundation even when the provider choice changes later.
Use transcript completion in downstream processes
When transcript handling is enabled, teams can treat transcript completion as an operational step in a larger workflow. The Completed Speech Transcript workflow event starts when a speech transcript is completed. This is useful when spoken interactions need review, escalation, follow-up actions, or reporting after processing finishes.
Add live speech-to-text to a voice experience
A provider can also turn speech into text while a person talks (realtime speech-to-text). Speech ElevenLabs provides this. It transcribes only; it does not reply. A voice assistant can use it as the listening step of a chain of speech-to-text, a chat model, and text-to-speech. See AI Realtime Voice for how this compares with native speech-to-speech models.
Operational Notes
- This module is the shared speech base, not the provider-specific configuration screen.
- It supports transcript handling, speech-related workflows, and secured access to redacted recording downloads when those features are used in the rollout.
- Provider-specific voice choices are managed through the related provider pages, not through this base page.
- Playing a redacted recording requires the Download redacted speech audio file permission.