by Maria Simonton, Director of Product Marketing, Interactive Northwest, Inc.
[frame src=”https://www.interactivenw.com/wp-content/uploads/2015/09/Speech-HiRes-300×300.jpg” link=”” target=”_self” width=”300″ height=”300″ alt=”Demystifying speech ” title=”Demystifying speech ” align=”right” prettyphoto=”false”]Since their early beginnings, speech recognition applications have often been viewed as expensive and complex—labels that slowed adoption amongst companies who were unwilling or unable to take the risks associated with deploying them. As the technology became more sophisticated and capable, misconceptions surrounding speech recognition continued to grow. Fast-forward to 2015, when speech interfaces are not only widespread, but expected. Increased offerings by a wider variety of vendors now permeate the marketplace, resulting in more architecture options, shorter deployment windows, and faster returns on investment. Never before has speech-enabling a self-service application been so attainable. By giving consideration to the following guidelines, all companies can deploy speech recognition in the contact center with confidence.
Identify Needs
All great speech applications begin with a great design. But before design comes purpose and intent. Companies must ask and answer:
- What problem are we trying to solve?
- What do our customers want and expect from an interaction?
- How will the solution benefit our users?
- How will we measure success after deployment?
One of the biggest mistakes that companies make when implementing a new self-service solution is to build to their own business requirements without taking into account the needs of the end user. A well-designed speech application should steer callers in a way that is natural, effortless, and—most importantly—relevant to their goals. Spending time upfront to analyze common user objectives is a valuable part of the process. Data can be gleaned from customer surveys, reporting from other self-service channels, and by speaking to contact center agents. Only after the heavy-hitter tasks have been identified can the application designer move on to the next question: Which transactions are best suited to speech, and which are not?
An effective speech design utilizes the technology where it will have the greatest impact. Understanding how and when—and even where—callers use the application dictates the style of speech interface that will be most effective.
Start Simple
Setting out to create a speech recognition solution that is all things to all users is an exercise in futility. Many companies try to get the most bang for their buck by cramming as many features into their application as possible, but this typically results in an overly-complicated experience for users. Furthermore, internal owners of the solution can be overwhelmed by the responsibility of designing, testing, and maintaining a large and unwieldy self-service application.
A more manageable way to implement speech, especially for first-time deployments, is in small, task-oriented chunks. Simple call routers are a great way to introduce speech recognition into the contact center environment. Additional capabilities can later be added to perform functions that alleviate agent handling and leverage the benefits of speech recognition, such as collecting a policy number or a payment.
Another good introduction to speech recognition technology is via IVR products and packaged applications. Products have a much shorter time-to-deployment phase, are thoroughly tested and hardened, and may even come bundled with speech licenses.
Engage the Experts
At the heart of an effective speech recognition application are its grammars, which serve as a blueprint for how the application behaves. A great deal of usability study goes into the creation of them, and a talented grammar writer must be part linguist/part psychologist to accurately anticipate what users will say. It’s precisely this type of expertise that distinguishes a serviceable speech application from a truly elegant one. Engaging an experienced speech vendor will go a long way toward making the process smooth and manageable from design through deployment, and ensure an exceptional user experience.
Protect Your Investment
The cost of developing and deploying speech recognition can often be justified in terms of hard and fast ROI numbers. While the application may pay for itself once a certain number of calls have been handled via self-service, a speech recognition application requires regular monitoring and adjustments in order to remain a sound investment—and to continue bringing returns. Business needs change over time, as do customer goals, and a speech application should be tuned periodically to keep it relevant and functioning well. A qualified speech vendor can guide companies down the path of application tuning and maintenance, ensuring that their initial investment holds value over time.
Schedule a consultation with an INI representative for more information on best practices in speech application development, or for details on INI’s speech-enabled product line.