We're excited to present Mist for IVR, our new TTS product specifically designed and tailored for IVR systems. We've developed Mist for IVR in close collaboration with our IVR customers to deliver the best-in-class quality for every type of business transaction over the phone. Let's see what this means.
The right model for the right use-case
At Rime, we're especially aware that not all speech is alike. The right style of speech for reading an audiobook is different from that of a casual conversation, and both of those are quite different from IVR.
Because of this, we are augmenting our current offerings. In addition to our state-of-the art Generic and Conversational models, Mist for IVR allows users to create professional, emotionally sensitive, and helpful voice experiences to delight and engage their clientele. Our architecture works with traditional, scripted text, and our fastest-in-the-biz latency allows us to integrate seamlessly with LLM generated text. Take a listen!
What makes Mist for IVR the best option for any sort of IVR use-case? We highlight a few features below.
Emotionally Appropriate Responses
Emotionally sensitive voice is extremely important. Without it, it's easy to inadvertantly insult your customers by not being correctly happy or sad or serious at the right momement. Mist for IVR comes with emotional appropriateness and sensitivity baked-in, so you don't have to worry about your brand coming across as flat, sarcastic, or worse. Take a listen:
Flexible Coverage for Novel Brand Names
One major hurdle for most TTS systems is the ability to pronounce novel words (like most brand names) and compound nouns (like, five piece meal deal). These are especially difficult when combined. But Mist for IVR offers out-of-the-box coverage for countless brand name compounds and our archictecture allows us to add new, bespoke coverage on-the-fly, because new products are being developed all the time. Below is an example of what we're talking about:
Turn-taking Subtleties
When conducting business over the phone, there is a lot of back-and-forth, and clearly indicating that the user was heard and understood is critical. To get this right involves numerous subtle speech moves to affirm that the information has been successfully communicated. This is easy for humans, but hard for TTS. Mist for IVR was designed to appropriately and consistently deliver this sort of confirmation, in the familiar ways: gotcha, mmhm, sure thing, etc.
Wrapping up
People speak in various ways, each dependent on the context and situation. And IVR is a genre distinct from casual conversation and narration. Mist for IVR will allow users to get the right performance to do the job right, delivered at scale and with our best-in-class latency. Give it a try today!