Many customers fear speech recognition IVRs -- a.k.a. voice self service -- and with good reason. There are many poor implementations that give a bad name to speech applications. Yet, many other good speech implementations continue to deliver unique and valued services to customers, and exceptional ROI for the companies that implement them. The goal underlying voice self-service implementation is not only to reduce cost, but also to improve customer service. Encouraged by the success of Web-based self-service, many companies see potential in applying the same key principles of Web self-service to voice self-service, such as expanded services, richer content, and user-driven search capabilities.
So how do you join the ranks of voice self-service exemplars? How do you overcome customers’ existing prejudices of speech recognition? The following top five keys can help you along the path to building superb voice self-service solutions.
1. Use “Deep Dive” ROI to Drive Your Requirements
The value of a robust ROI calculation cannot be overemphasized. However, many companies begin implementation without solid data available base for the calculation. One consequence is that the returns are underestimated, with companies overlooking key areas to generate returns. With a robust ROI calculation in place, these companies would have had a more profitable goal in sight for the project.
Being able to estimate the ROI will allow you to decide how to get the most value from a voice self-service investment. But to do a good ROI for speech, you have to go deep. You can’t merely calculate the number of incoming calls and multiply by a “SWAG” success rate. A reliable ROI will be based on the following data elements:
- The most common call types/call reasons
- The average handle time (AHT) for each call type
- The breakdown of AHT for each call segment (e.g., Greeting, Verification, Resolution)
- The estimated completion/containment rate with speech, based on detailed analysis of input types, transaction complexity, and data availability
- The estimated reduction in handle time for opt-out calls
This analysis can help you identify additional areas of value, often overlooked, and help you design requirements that will help you achieve your profitability goals. eLoyalty has proven templates to help you quickly calculate an accurate ROI.
2.Design Your Application for Real-World Usage
Good speech design is a combination of Linguistics, Art, and Science. For example:
- Linguistics - Determine for each prompt what is clear, what is intuitive, and what are the natural responses people are going to give.
- Art - Make sure that you have some clever people working for you to design a smart speech application. Look for ways to convey branding, guess customer information, and create more lifelike customer interactions to delight customers.
- Science - All possible scenarios must be designed for in the application. The decision tree does not only need to be complete, but it also needs to take all response options into consideration.
3. Integrate with Your CRM System
By integrating the speech application with your CRM system, you will be able to optimize the user-friendliness of the application while adding valuable information to your CRM data. CRM integration allows you to offer “smart” prompts in your IVR. For example, if a customer has not paid last month’s bill, you may give them a custom prompt that says: “If you are calling about a late payment, please say ‘Yes’, otherwise, say “Main Menu”.
4. Integrate with Your Contact Center
Opt-out calls are often seen as a result of poor or incomplete speech application design. This is not always the case, nor does it have to be. When integrated with the contact center, speech recognition can reduce call handle time for opt-out callers and create favorable customer experiences and a higher re-use rate for your speech applications.
5. Test the Application Completely
When testing speech application, take into account: performance, scalability, and usability, both from a voice quality and speaker independence perspective. With a thorough test of all application pieces, you can ensure that the application works in the way it was designed, and also ensure the usability of the application. Pilot and re-test. Typically, you will make smaller adjustments in the test phase that will improve the usability of your application significantly.
About the Author
Mike Ashe
Mike Ashe is a Vice President with eLoyalty, where he leads the global Contact Center Optimization Solutions service line. Mike has worked with hundreds of companies over the last 20 years, with a strong focus on keeping customers happier (and loyal!) while reducing service costs. He has worked and done analysis in over 70 contact centers in nine countries around the world. Key clients include HP, Dell, Avaya, Agilent, Bank of America, Wells Fargo, many Blue Cross/Blue Shield companies, Kaiser, Toyota, Nissan, Disney, BP, T-Mobile, AT&T, American Express, and GlaxoSmithKline.