When building a voice application one of the first questions you need to answer is: to what degree will you use text-to-speech? Initially, it may seem like a great idea to use text-to-speech (TTS) wherever possible. After all, it’s faster and cheaper to just type out some generic code, right?
Here’s the catch. TTS is great! For some things… If you have variables that change for every caller, things like an ID number, an account balance, or even timestamps, these are a great fit for TTS. The reasons for this are two-fold:
- These are typically short phrases, and
- Callers have an expectation for this type of data.
For example, if a caller needs to pay a bill they expect the TTS engine to provide a given value in dollars and cents. The overall result here is a streamlined call with little to no confusion for the caller.
Potential TTS Pitfalls
On the flipside, rendering an entire call-flow in TTS violates both principles of brevity and expectation. Using too much TTS has significant consequences:
- Lack of clarity: It can be difficult for end-users to understand long passages of TTS. This means callers repeat menus more and call times increase. Even if you have a good TTS engine and reliable telephony backend, like what Plum provides, you can’t control your caller’s environment.
- More transfers: The whole point of automated customer self-service is to increase efficiency. But frustrated customers that don’t just hang up are much more likely to transfer out of your IVR to speak with a live agent. The net effect of this is to drive up your customer service costs.
- Bad customer experience: All of this adds up to a poor customer experience. In a day-and-age where customer service is a major differentiator, companies can’t afford to sacrifice customers at the altar of bad customer experience.
So, unless you plan to send ear trumpets to all your customers to help them hear the menus better, it’s best to keep the TTS to a minimum.
Use Pre-Recorded Prompts
How, then, do you avoid ruining customer experience with too much TTS? Use pre-recorded audio prompts for static information in your call-flow. This includes things like hours, contact information, address, or instructions for customer transactions.
Having professional voice talent record your audio prompts is not a costly endeavor. Plus, the extra clarity and natural human cadence of pre-recorded prompts help to temper customer frustration and its aftermath. If you’re unsure where to find good voice talent, your IVR vendor should be able to provide some direction. Here at Plum, we provide consulting services to help customers navigate this process.
Having a completed call-flow kills two birds with one stone. Not only does it provide you with an overview of your entire application. But it also provides the script your voice talent needs to record the prompts for your application. So, while the time and energy required to set up pre-recorded prompts may seem high, in reality, it’s a fast and cost-effective approach that dramatically improves the quality of your voice applications.