Migrating From Touch-tone to Speech: Making a business case for migration development
The development of speech technologies is creating increased opportunities to leverage speech resources across multiple business solutions. A well-defined service approach to speech design in combination with a comprehensive speech technology strategy is essential to success when migrating from touch-tone to speech recognition. Historically, Interactive Voice Response (IVR) applications have been implemented as stand-alone solutions. A comprehensive approach anticipates future integration opportunities and fosters incremental growth.
An effective business case for migration to speech provides a number of significant benefits:
- It defines the voice commerce opportunities available;
- It specifies which of those opportunities are not realizable with touch-tone service; and
- It examines the future evolution of the application.
The business case for migration to speech defines the voice commerce opportunities available, which are not fully realized with touch-tone service. A well-defined business case also provides input to guide developers in shifting knowledge from analysis to design. This translates into future flexibility for the application interface and business rules so the application can evolve based on changing customer needs and trends for new product solutions.
Voice Commerce Opportunities
As part of an enterprise-wide e-business strategy, benefits routinely associated with automated phone service are cost-savings, consistency, increased customer satisfaction, scalability and the ability to deliver solutions for repetitive inquiries efficiently. Other key voice commerce opportunities are reinforcing a competitive position in the marketplace by offering comparable services and projecting a business persona via branding strategies to reinforce the relationship and evoke a positive emotional response. Speech applications can project their brand through the use of music, a recognized voice or slogan, a personality or style and presentation tone more dynamically than touch-tone service.
Additional opportunities include optimizing operating revenue through higher system utilization, revenue generation strategies like cross-selling, potential operating cost reductions and enhanced customer service levels. Also driving the business case is creating a more robust voice platform in preparation for future development of Voice over Internet Protocol (VoIP), voice authentication, or adding complex functions that aren’t customer friendly in a touch-tone environment.The enterprise objectives must be specified prior to the determination of the unique blend of services to include with speech. These services may include expanded client environments, new target audiences, marketing strategies or aligning speech options to mirror other service channel offerings. It is important to look at function automation from a customer-centric approach and ask the following questions: How important is the outcome of the function to the customer? What is the level of function complexity? What is the customer’s anticipated emotional state while using the function? Is the customer more concerned with privacy or human contact during the function?
The customer-centric approach, as it relates to this computer-human interaction, defines the process of utilizing end user preferences to guide application design and development. The preferences of the target audience are integrated with defined voice commerce opportunities and service channel requirements. Continued development of the business case effectively strikes a balance between the preferences of the users and the economic feasibility of the opportunities.
Economic Feasibility and ROI
The economic feasibility analysis of function automation begins with the definition of basic variables for each evaluated function. The customer identification variables include the customer type, demographics and technical sophistication. The service level variables include expected response times, peak utilization periods, transaction rates, anticipated growth rates and the initial daily volume estimates. Information requirements are variables, which provide a basic description of the transaction, the type of data required and the relative size of the transaction.A return on investment analysis determines which functions to be included in the final design. For each function this analysis includes a statement of assumptions, constraints, and a comparison of function alternatives. The ROI feasibility analysis includes attachments defining the calculation methodology and all reference documents. The ROI basic formula is the cost reduction plus the revenue increase, divided by the total cost of ownership. Intangible factors like brand recognition and customer loyalty are factored into the basic formula through weighting techniques. Finally, tangible benefits are defined by asking does this function support the enterprise service channel goals and objectives? This process produces a comprehensive estimate of future ROI.
When customer services are designed to maximize the potential of the speech application, the ROI is optimized. Productivity and capacity are two customer service factors that influence the projected ROI benefits of the planned speech environment. Productivity increases when the time required for function execution is reduced while capacity is the potential to increase the number of times a function is completed in a defined time period. These service factors are evaluated by identifying discrete function tasks, completing a task/time analysis and associating a skill level with each task. Finally, the financial breakeven point, which ideally is measured in months, is calculated.
After calculating the ROI, it is important to take process snapshots reflecting anticipated pre- and post-implementation metrics. This will facilitate post-deployment measurements used to fine-tune the speech environment to achieve the projected ROI. When possible, replace calculation assumptions with quantifiable cost, benefit and risk data generated from company experience. This will increase the reliability of the ROI modeling and success factors used to evaluate the speech deployment.
Customer Usability Metrics
Quantifiable business goal metrics like system performance and service level standards are now integrated with customer metrics. In a successful speech recognition environment, companies leverage the personal characteristics of the computer-human experience by first capturing the customer interaction, identifying how the needs were satisfied and interpreting the intent of the contact. This defined set of user metrics is then used to assess, monitor and manage the customer experience.
Before defining customer usability metrics, it is necessary to understand the scope and attributes of the intended audience, and consider the following. Is the target audience an existing voice response audience, or a new audience? Will the demographics of this audience change over time, and if so, what is the frequency and nature of the change? Is the focus of the application primarily external-call center and self-service functions, or internal-employee resources? Can a single speech environment meet the needs of all audience members, or will the audience require multiple speech environments for delivery of customized experiences? The voice of the customer must be captured and interpreted. Some reactive channel tools are customer complaints, service hot lines, technical support lines and increased system usage statistics. Proactive tools like internal/external focus groups, surveys providing user feedback, prototype usability testing and system benchmarks also provide valuable insight. Performance benchmarks can be defined using system utilization reports, market penetration reports and system performance reports. I have found that two key metrics are the look and feel of using the system, and the learning curve, which is defined as the number of repetitions required for a novice user to become a power user.
In addition to the outcome of the function, customers measure system usability throughout the interaction when information or assistance is required to continue. Customers demand more from the interaction than just information. They also measure their experience in terms of responsiveness, a measure of both time and the quality of vocabulary recognition. Optimizing this interaction requires a thorough analysis of customer usage metrics. Will the customer typically call from a cell phone or use a speaker? Does the customer usually call from their home, office or a car while traveling? While speech technology has evolved, the service expectations of the customer have remained the same. With this in mind, the speech experience is engineered to meet the expectations of the customer, which in turn will substantially boost system productivity.
Defining the Speech Environment
The challenge for these no-nonsense times is to create a speech environment, which is equally focused on ROI analysis, customer metrics and voice commerce opportunities identified in the business case. The ability to leverage the design as part of future solutions is also critical. The business case provides the inputs and best practices for this intuitive design.
The economic feasibility analysis also includes defining the scope for the functions to include in the speech environment. A broadly defined scope includes multiple customer environments and functions. Therefore, there is an increase in design complexity, cost, development time, implementation benefits and time required for system tuning. The benefits of employing a narrower scope include valuable lessons learned in the initial deployment and decreased development time. Speed is of the essence because windows of opportunity open and close continually. A flexible design permits companies to recognize and move rapidly to exploit changes in the competitive market environment.
Additional variables for the migration from touch-tone to speech are now defined which contribute to a successful integrated service model. What is the database, operating system and data services requirements that will permit seamless integration of business rules and data? Do customer expectations require continuation of touch-tone functionality, generating considerations for the cost of maintenance, management resources and future consistency between speech and touch-tone? How will the business assist customers with the new speech environment? This includes system design elements like help prompts, menu design, re-prompts, and opt-out features; as well as marketing and customer communications. Will customers require a multi-lingual speech environment? The business case customer-centric approach gives customers a vested interest in the speech environment design, which enhances customer loyalty and strengthens business relationships. k
John Cootz is senior project coordinator at AIG VALIC. He can be reached at john_cootz@aigvalic.com.