Get the caller engaged as quickly as possible
Callers will completely zone on you if you don’t get them involved quickly. If they sense at the beginning that the system is playing fluff to them that’s not important, they tune out. Getting them back is difficult if not impossible. Get them involved and answering a question or providing information as quickly as possible.

Skip telling them what you can help them with, or what the system does, or how wonderful it is. Either the system can help them or it can’t. Give them their options and let them drive. (See What Not to Include at the Beginning for more on what to skip.)

Identify the IVR they have reached
Give the caller a brief description of the IVR. For example, “Welcome to the ABC Corporation Bill Payment system”. Callers who believe they have reached the wrong number will immediately hang up.

The very first message should be short and simple, making it clear that the caller is talking to a system rather than a person, but without saying this explicitly (which would take time and probably distract the caller). Analyses of the interactions of human agents with callers indicates that human agents rarely start the interaction with the word "Welcome" (Lewis, 2011), so that works well as the first word in the introduction of an IVR (e.g., "Welcome to AutoRez car rentals").

While we want to eliminate fluff, the IVR is a customer touchpoint and an opportunity to reinforce a brand with callers. Make it obvious whom the caller has called, and give the IVR as much presence as a web site. Keep it short, but play some kind of earcon that callers will associate with the brand from marketing campaigns, if available. A short tagline (5-10 words max) can also be employed here, and optimally should be refreshed with new marketing campaigns. An example: “Welcome to AT&T <earcon> Rethink Possible. How can I help you?”

Craft the introduction carefully
The key to rapid engagement is in the crafting of the introduction. Unlike any other part of the application, 100% of calls will go through the introduction, so it's critical to have an introduction that engages callers as quickly and concisely as possible (Rolandi, 2004b, 2007a; Suhm, 2008). In addition to losing the caller's focus, time is money in the IVR. "Every unnecessary syllable in an introduction to an often-used IVR can cause expense for the enterprise and wastes the caller's time" (Lewis, 2011, p. 201). Keeping in mind that the costs of IVR operation can vary considerably from setting to setting, Yudkowsky (2008) reported a $1 million per year savings for AT&T for each second of speech removed from a presumably high-volume IVR.

Don't engage too quickly
On the flip side of engaging rapidly, if the greeting and first question are super short, callers haven’t necessarily shifted their focus from whatever they were doing before. How short is too short hasn’t been quantified, but it has been observed, so watch out for it.

Make the first question an easy one
It's always a good idea to start the dialogue flow off with a question that's easy for users to answer. This isn't always possible, but when it can be achieved, it puts the user at ease and shows her that she can be successful using the system. Starting off with an obscure question that may require the user to find a piece of information from a statement, for example, may be frustrating to the user and affect her success at using the system.

Pause after the welcome message in high touch applications
If most callers use the application on a frequent basis, consider pausing for about a second after the first Welcome message (and before any prompting) so experienced callers (who already know what to say) can barge in comfortably (Balentine & Morgan, 2001), then continue with the rest of the introduction. Do this only if you're sure that the time savings due to rapid barge-in will compensate for the extra silence playing when callers do not barge in.

References

Balentine, B., & Morgan, D. P. (2001). How to build a speech recognition application: A style guide for telephony dialogues, 2nd edition. San Ramon, CA: EIG Press.

Lewis, J. R. (2011). Practical speech user interface design. Boca Raton, FL: CRC Press, Taylor & Francis Group.

Rolandi, W. (2004b). Rolandi's razor. Speech Technology, 9(4), 39.

Suhm, B. (2008). IVR usability engineering using guidelines and analyses of end-to-end calls. In D. Gardner-Bonneau & H. E. Blanchard (Eds.), Human factors and voice interactive systems, 2nd edition (pp. 1-41). New York, NY: Springer.

Yudkowsky, M. (2008). The creepiness factor. Speech Technology, 13(8), 4.